Wednesday, September 25, 2013

Reducing Retirement Risk with a Rising Equity Glidepath

Michael Kitces and I have finished a second research article together. This one is called, "Reducing Retirement Risk with a Rising Equity Glidepath." Tara Siegel Bernard first introduced the article in the September 14 issue of the New York Times, in a column cleverly named, "An Adage Adjustment for Investors at Retirement."  Today, Michael and I are each 'introducing' the article with posts at our respective blogs. Michael's detailed post on the article can be found here

As for my post, this is part of a new experiment. I recorded five video blog entries with cooperation from The Wealth Channel at the American College. So my explanation of the article is provided in video form:





For email readers, the videos never show up in the email, but you can see the video by clicking here.

11 comments:

  1. I found the content interesting, however the music in the background was very distracting. Maybe you could use the music for intro and at closing, but definitely "lose the music".

    ReplyDelete
    Replies
    1. Thank you. I will pass your comment on to the editing team, as I was not involved in that process.

      Delete
  2. Wade:

    I‘ve done a simple experiment that suggests desirability of rising equity allocation depends on how good results have been in years-so-far. It agrees with your-and-Michael’s conclusions for “good” or “bad” early years -- but suggests there’s a middle range of early-years results for which it’s better to hold equity constant or reduce it for years ahead. I’ll send illustrations to you and Michael by email.

    The reasoning is this:

    If early-years results are “bad” or “good,” in either case the dominant consideration is equities’ better upside possibilities – if “bad” you need upside, if good you don’t worry about equities downside and want their upside. Just as you said.

    But in the middle range, what becomes more important is the downside possibilities, and for those more equities is worse.

    One way to deal with this would be to, through testing, develop year-by-year rules for response to result-so-far: if below X or above Y, increase equity, but if between X and Y hold or reduce equity. Another way would be use of dynamic programming which has a similar effect.

    Dick Purcell

    ReplyDelete
    Replies
    1. Dick,

      This sounds reasonable.

      I must emphasize that what Michael and I discuss is more of a default starting point, or a glidepath which could be used by the lifecycle asset allocation / target date funds. This is for individuals who do not want to dig in to personalized dynamic asset allocation strategies.

      But perhaps what you are talking about is the funded ratio, or the ratio assets to the present value of future spending needs. In simulations where the funded ratio stays close to 1, then gradually raising the stock allocation may not be a good idea. Whereas, the parts of the distribution that we investigate with are results are the downside worst-case type scenarios (where funded ratio is less than 1, and well as the median scenarios where a 4% withdrawal rate will allow the funded ratio to get much greater than 1. We didn't look much at that middle ground where the funded ratio stays near 1. It's worth looking at more, though then, of course, we are moving toward the dynamic programming type of approach instead, rather than looking for a reasonable default glidepath for those to use who do not otherwise want to pay much attention to this issue.

      Delete
    2. Wade –

      Thanks for your reply.

      But what I’m suggesting does not require dynamic programming, nor does it require continuing analyses during aging after retirement. Based on some testing, you could at age 65 set POLICY for all future years that determines each year’s allocation based on balance at that time relative to future goals. Could be as simple as the X, Y rules I mentioned in my first comment.

      Frankly I think it is at best premature to be spreading all this whooping about superiority of rising glidepath based on good and bad extremes. Further testing MAY reveal that for a large percentage of people, in a large percentage of years, they may be in the middle range where increasing equity is not best.

      Dick Purcell

      Delete
  3. Wade
    For a 40 year retirement with 90% certainty of not exhausting funds is 3% a reasonable SWR from your data. This assumes a starting euqity allocation of 30% rising to 60% equity.

    ReplyDelete
    Replies
    1. It depends on the capital market expectations. With what we think is most reasonable, a 2.7% withdrawal rate matches your specifications. But with historical averages guiding market simulations, the withdrawal rate is 3.7%

      Delete
  4. In your journal article you mentioned that the results were similar for a 40-year time horizon but didn't share that data. Do you have any tables you can share on success rate for 3% withdrawals, shortfalls at 5%/10% distribution, etc. and the results of different glide paths? Thanks.

    ReplyDelete
  5. Dan,
    I thought those would be in an online appendix, but I see that they weren't included.
    I'll try to remember to get those into a file which can be posted online at some point.
    Thank you, Wade

    ReplyDelete
  6. So how is one supposed to start retirement w/ only a 30% allocation in equity? Are you suggestiong they rebalance at the start of retirement to likely a much lower equity allocation of 30% and then gradually increase that equity allocation through retirement? Or are you suggesting that one should decrease their equity allocation as they approach retirement and then increase it again while they're in retirement? Ty

    ReplyDelete
    Replies
    1. The latter. No big jumps. Lifetime stock allocation would follow a U-shaped pattern.

      Delete