Thursday, November 1, 2012

Two Schools of Thought on Retirement Income

The following figure is of a table which compares and contrasts the probability-based and safety-first schools of thought regarding retirement income. You can click on it to see a bigger version. I could not get this formatted to appear directly as a table embedded into the blog:


  1. Wade

    I’m a financial planner in Australia. The approach to retirement planning we currently take incorporates many aspects of lifecycle finance theory (e.g. stock risk, human capital) but there are others we find hard to embrace.

    A clear one is the role of inflation-adjusted single premium immediate annuities. In Australia, lifetime annuities are prohibitively expensive, due to a lack of active providers and appropriate hedging vehicles. We are uncomfortable covering “basic needs” at any cost.

    More arguably, my observations are people are not as rational as life cycle theory assumes. Of course, the experience of the past 5 years has seen a flight to safety and the appeal for “locking in” basic needs has risen, together with the interest in the lifecycle approach.

    But as some of the behavioural economics research suggests, I suspect people are more driven by relative rather than absolute goals. They want to be better off, or no worse off, relative to the people they compare themselves with.

    I doubt whether I will be thanked by a retiring client in 10-20 years time for pandering to his current need for “safety” and expensively covering his basic needs now, when it turns out that he ends up worse off than those he compares himself with, despite achieving his absolute financial goals. And history, for what it’s worth, suggests this is the most likely outcome.

    While lifecycle finance theory has added a lot to our thinking on retirement planning, its underlying assumptions don’t accord with how people actually behave i.e. they aren’t rational agents that optimise their behaviour to best satisfy their preferences. Therefore, it’s not surprising that its recommendations don’t always sit comfortably with some financial planning practitioners.

    John Leske


    1. John,

      Thank you for your thoughtful reply.

      That is an interesting point you make about absolute vs. relative goals. If this motivates behavior, then missing out on some of the upside potential enjoyed by one's peers by focusing on a safety-first approach could be a "painful experience" that would serve as a real-world explanation for why people tend toward the probability-based approach.

      Thanks, Wade

    2. Hi John & Wade,

      As an Australian adviser as well, I share your frustration with the lack of annuity providers and options. The Federal budget last May should hopefully see solutions emerge in the next 1-2 years.

      In saying that, my perspective is a lean towards the 'safety-first' approach that Wade has summarised above quite well.

      Compared with the USA however, we do have the additional benefit of the Aged Pension system in Australia and particularly when complemented with an annuity type income stream, it can assist when addressing longevity risk and a floor of minimum income for certain retiree demographics. If nothing else, it does provide a contingency benefit.

      As for the real-world explanation of probability based popularity, what I do tend to see is the 'adviser risk attitude' is quite often reflected in a high majority of client's portfolios. So if they have a strong comfort with equities, then this is typically the client's outcome for retirement income. The probability vs safety first aspect for popularity then really tends to be more of an adviser-driven outcome for the client. This has the potential to change in future years as advisers may become more focused upon retirement outcomes that perhaps are not heavily reliant upon growth asset classes.

      What I find most interesting are the significant amounts of research that indicate retirees are seeking safety of capital above that of a return on capital. This is common for all spectrums of net worth. Yet from my experience, most client's portfolios are based upon the probability-based method.

      A couple of key reasons that I am now moving towards the safety first approach is that the market research indicates that this is what many retirees are seeking first and foremost. The discussions with the client are then not positioning the adviser as a spokesperson for the equity market, or if 'x' occurs, then 'y' will be the result, but rather a broader discussion and application of capital so as to secure a smooth monthly income and provide peace of mind.

      The second reason is one of the most debatable. From my understanding, probability based portfolios are reliant upon reversion to the mean, particularly for equities. Again, my understanding is that several academics have proven that over the long term, there is no statistical evidence that this is a guarantee that equities would revert to their mean.

      So the resultant view of mine is that the combination of a demographic that are seeking safe income not associated with more volatile assets, complemented by research that those riskier asset classes are not necessarily guaranteed to revert to their mean, are a couple of key reasons why I see merit in the safety first approach for retiree clients.

      Thanks again for the articles Wade.

  2. I believe that in your table you have ignored the most important difference between the two approaches - namely the goal of each approach.

    In the safety first approach the goal is to meet the aspirational level of retirement income with a fairly high level of probability and a side constraint of meeting a floor level of retirement income with near certainty.

    The goal of the probability based approach is fuzzy - but often is a level of financial wealth at retirement, hence the SWR strategies. This approach to retirement income is problematic for a host of reasons - not the least of which are the following two. It attempts to marry volatile portfolio movements to a smooth level of retirement income. It ignores the level of current real interest rates in planning for retirement in both pre-retirement and during retirement.

    1. Thank you. These are some good points that should be better incorporated into the table!

      I agree with your safety-first description.

      For the probability-based approach, I think it is essential to meet your lifestyle goal (basic needs + aspirational goals) with a high degree of probability as determined by the unlikely failure when using a safe withdrawal rate. I fully agree with the problems you mention.

  3. Wade - the table is very helpful in general, but also helps explain why I sometimes feel a little schizophrenic when thinking about how I help clients plan for retirement. I have feet in both camps. Where would you put yourself? Or are we going to see a 3rd column evolve?

    1. Kathy,

      Thanks. Yes, I bet having feet in both camps is common. I feel like I'm now 2/3 safety-first and 1/3 probability-based. I don't know if there will be any third column, but I do hope to see more and more efforts to better integrate the two approached together and arrive at some common understandings.

  4. Wade, I love this table, but I needed to read no further than the first row.

    I would add Bernstein, Kotlikoff, and Sharpe to the right column and then wonder, with all due respect, why anyone would think this is a competition.


    1. Thanks Dirk.

      Well, not everyone trusts those academics with their heads in the ivory towers. But good suggestions on adding those names.

      There are also other practicing financial planners on the safety-first side (in addition to yourself) such as Paula Hogan and Rick Miller.

  5. This "stocks always risky in the long run" idea bugs me. The implication, rising even to explicit admonitions from people like Bodie, is that individuals saving for retirement should not therefore invest in stocks. Ok, where does that leave pension funds? If they are supposed to match liabilities and assets, aren't they being rather foolish when, as they all do, they invest heavily in equities? And by the way, if pension funds took the advice and invested only in safe TIPS, where the heck would equity capital come from?

    Another thing in the safety-first column makes me wonder - if current market conditions matter then doesn't the relationship between bond returns and stock yields in the USA kind of suggest that bonds are rather risky at the moment with negative real returns while stocks are priced for modest but positive returns?

    1. Thanks for your comments, sorry for taking so long to reply.

      You make good points and I appreciate your taking time to share!


  6. I have one word for both columns; Scientsim