Hi Guys,
If you don’t know, I have some skills and experience that contribute to my qualifications in assessing Monte Carlo-based retirement analyses. I have reviewed much of the literature, have done countless specific analyses, and have even written a Monte Carlo code when not many existed. So, this is not a casual posting; it has a serious purpose.
The NY Times article that reviewed the Pfau and Kitces study did a respectable job at summarizing the researcher’s basic findings. But, like any broad-brush evaluation aimed at a general readership, it is not necessarily nuanced or complete. A study of this magnitude deserves special attention since an interpretation of its discoveries can be very personal. Also, the authors are heavyweights in the retirement arena; these researchers speak with authority.
Therefore, my first recommendation is that you not be satisfied with a reasonable review, but that you go directly to the source. The referenced document is not mathematical, is a breezy read, has specific recommendations, contains useful charts that summarize all its findings, and is only19 pages long. Here is a Link to this fine Monte Carlo research paper:
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2324930The study itself is comprehensive, but it is simultaneously limited in scope because it only partially deals with the complex continuum of retirement issues. For example, retirement planning must address both the accumulation and the distribution phases of an overall retirement strategy. The referenced study only focuses on the distribution portion of retirement.
The study examines the survival prospects of a retirement portfolio under the commonly accepted 4 % and 5 % annual withdrawal schedules (adjusted for inflation) for three representative market rewards scenarios. Not shockingly, the postulated returns scenarios are a primary determinant force in any portfolio survival study. The three reasonable postulated scenarios are: one developed by Harold Evensky for professional financial planning purposes, another calibrated to the current low interest rate environment, and a third that reflects historical equity and bond returns.
The major distinction in the Pfau and Kitces work is that the equity/bond asset allocation mix is not held constant during the retirement period; 121 equity glide-pathways are defined and evaluated. Monte Carlo analyses randomly selects the portfolio’s returns annually for each of 10,000 cases considered for each equity glide-path. A 30-year retirement lifecycle is documented.
In some instances, the retirement portfolio is favored with positive returns in its early years; in other instances, the reverse is randomly selected and the portfolio suffers initial erosive market drawdowns. Portfolio survival is definitely dependent upon both the magnitude and the order of these future projected returns.
For a more complete assessment of the entire retirement planning process, I recommend you visit another research paper. The paper is authored by Javier Estrada, another financial research wizard; his paper is titled “Rethinking Risk”. Here is a Link to the compact 23 page study:
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2318961Estrada examined returns data from 19 countries over a 110 year timeframe. He makes the case for a heavier commitment to equity positions during the accumulation phase of retirement planning, even just before the retirement date. He observes that “ In fact, stocks have both a higher upside potential and a more limited downside potential
than bonds, even when tail risks strike.” He’s writing here about Black Swan low probability events late in the lifecycle period.
In his conclusion section, Estrada writes “ … it is clear that stocks are more volatile than bonds, but it is far from clear that they are riskier than bonds for the type of investor considered here. This is because, even when tail risks strike, stocks enable investors to accumulate more wealth by the end of the holding period than bonds. Hence, in what sense are stocks riskier than bonds for a long‐term investor that focuses on the endgame?”
In essence, the Pfau & Kitces Monte Carlo studies, and the Estrada empirical assembly and analysis of market data dovetail to fit an emerging picture. These researchers are advocating for more aggressive portfolios with a higher fraction of equity holdings. However, Pfau & Kitces endorse a U-shaped percentage equity holding profile that features more fixed income products immediately after the retirement date.
Their study is soundly constructed, but its numerical findings are not overwhelming. That’s why you need to examine the results for yourself. It is likely that some of the reported trends are within the uncertainty (the noise) of the calculations. Basically, the study documents marginal benefits, small
gains in survival likelihoods.
In many instances, the reported portfolio survival probabilities are at totally unacceptable survival rates. That is especially true for the Evensky model returns and today’s low fixed income returns forecasts. From my viewpoint, the only actionable portfolio equity glide-path scenario is that coupled to the historical market returns. That’s a rude awakening.
Referring to the Pfau & Kitces figures, their analyses show portfolio survival rates above the 90 % likelihood mark only for historical-like annual return rates. That might also be interpreted as a clue that a 4 % withdrawal rate is unacceptable if current, muted market conditions hold for the next decade or so.
There is a lot to be learned from the two referenced studies. Please take advantage of them. I believe that although they offer interesting and new insights, they do not make overwhelmingly compelling arguments. If I do decide to act on them, it will be very incremental in character.
Nobody sees the future market returns with clarity, so conservative retirement approaches and planning are always the order of the day.
I hope these references are helpful in that regard.
Best Wishes.
