Monday, November 12, 2012

Video Interview: Choosing a Retirement Income Strategy

In August, I recorded the following interview for the New York Life Center for Retirement Income at the American College about my recent research paper, "Choosing a Retirement Income Strategy: A New Evaluation Framework." This article was published in the Fall 2012 Retirement Management Journal, and it was also a co-winner of their Academic Thought Leadership Award (along with "Optimal Withdrawal Strategy for Retirement Income Portfolios").

In the interview, I explain about how I've been thinking to evaluate different types of retirement income strategies. The interview includes some additional graphics about how each strategy works, which are not found in the paper, as well as analysis about how the strategies perform using a variety of outcome measures.

This interview will be part of the American College's Retirement Income Certified Professional (RICP) designation program. They are building a huge collection of thoughtful interviews and it will be great to see these rolled out publicly over time. Already, there is a lot of  good content freely available on their website. The interview below is not yet on their website, but I thank them for providing me with the opportunity to post it here.


  1. A very nice presentation. One comment (as a retired professor): It is counterintuitive to use a measure of changes in income over time constructed so that decreases give high values and increases give low ones. It would be easier to understand if you did final/initial, so high values are increases and low values are decreases, instead of initial/final.


    1. Thank you for the comment, and yes you are right. That would make the measure easier to understand!

      I didn't do that, because then it would be the exception that with all of the other measures, higher was better.

      Nonetheless, the way I defined the measure confuses me as well, as each time I want to use it I have to think through the steps. Thank you.

  2. Interesting stuff.
    The key issue: the intent and usefulness of the Monte Carlo simulation is surely to gauge the extremes or how bad it could get for the retiree(s). So the assumptions of the Monte Carlo matter a lot. Are the returns assumed to be normally distributed or are fat tails modelled? How does the inter-dependence of stock and bond returns along with inflation get reflected? I cannot believe the economy randomly and independently assembles combinations of the three. Things often go bad in bunches. If high consistent/hyper inflation hits, will real-return SPIAS be honored? The promise of a GLWB is only as good as the guarantor. Correlations themselves have ranges. Things get too complicated.

    The original stress testing by the constant withdrawal method was simply to use actual history, in Bengen's case, only the USA. Therein lies the crux of the matter - how much will the USA continue to be the USA, or Canada (in my case) continue on its past patterns of variations and trends. Your previous research on safe savings rates for international countries showed some countries like Germany, Japan, Italy, Finland, Belgium experienced impossible worst cases for stock/bond savers. Would any of the guaranteed options in your retirement income menu have done any better in the times of economic calamity that caused those situations?

    Re the systematic withdrawal case, most investors managing their own portfolios have been shown to under-perform the asset class returns by making poor investment decisions, so maybe a scenario adjusting the returns downward would be useful. It's a qualitative argument in favour of at least some annuitization too.

    1. Hi,

      Thanks for the questions. The return distributions are multivariate lognormal, taking into account the correlations between the real asset returns and inflation. However, there are not fat tails, so things could get worse.

      You are right to point out that really the distributions should be wider.

      About the viability of annuity providers, I have to defer to the opinion of Joseph Tomlinson, who is someone I trust.

  3. FYI.

    Your video seems to not be loading

    1. Thanks, I checked and it seems to be working now.