Putnam Optimal Asset Allocation Review

Reader Question

I recently read Putnam's optimum allocation for downside risk and they say 10 to25% in stocks, but they don't give you the allocation on stocks. I like to get your evaluation of this paper.

My Reply

Thanks for writing. The paper in question is, I assume, Optimal Asset Allocation in Retirement: A Downside Risk Perspective published by Putnam Investments (a big mutual fund company) in June 2011. This study attempts to suggest the most appropriate portfolio asset allocation to last its owner's expected lifetime with a set withdrawal rate (such as starting with 6% of portfolio in the first year and then growing with inflation).

The paper is interesting and I see no reason to question its conclusions given its assumptions. But its assumptions are very much out of touch with reality for today's individual investors. That makes the conclusions be of purely academic interest. I would certainly not plan my retirement today based on the numbers suggested by the paper.

I will outline the two biggest problems with the paper's assumptions and one interesting nugget of information presented in the paper that I think is relevant for today's retirees, even given the flaws in the rest of the study.

What Happens If You Live Longer Than Expected?

Skimming through the paper and its tables you immediately notice that it seems to analyze annual withdrawals that start at 5-8% of the portfolio and go way up from there. How can that possibly square with conventional advice to not withdraw more than 4% of the portfolio if you wish to see it last 30 years? It squares because Putnam paper is not concerned with the 30-year horizon. It uses Social Security Administration's Actuarial Life Tables for expected lifespans and bases its withdrawals rate on corresponding horizons. For example, today's 65-year old male is expected to live on average another 17 years. Accordingly, the study's conclusions are optimized for that 17-year span. Of course, the actual Monte Carlo simulations Putnam used to come with their conclusions considered all possible timespans from dying the same year to living to be 110 assigning each possible age an appropriate probability. But the net result is still that Putnam's conclusions are appropriate for investors who live as long as SSA's actuarial tables suggest.

This kind of analysis -- planning to run out of money at right around the expected time of death -- is very appropriate for annuity companies or pension funds, which is why the biggest pension fund of all -- the Social Security system -- has those actuarial tables handy. Such entities have thousands or millions of beneficiaries and such large numbers all but guarantee that an average 65-year old male beneficiary will in fact die by the time he is 82. That makes for relatively easy and reliable planning and is a large part of the reason why the SSA can fairly accurately project its surpluses or shortfalls decades in advance. But such logic is not appropriate for an individual investor planning for him/herself and the spouse.

This is a big problem for individual investors when taking Putnam's suggestions at face value. What if you happen to live longer than an average person? Putnam's very conservative portfolio recommendations are optimized to last for an average person's lifetime. What will you do when you run out of money at 82 after starting withdrawals at age 65 with a very cash-heavy allocation advised by Putnam and spending 6% of your portfolio the first year and increasing those withdrawals in subsequent years? What if you are destined to live till 90? Your final years -- when, arguably, you need that money the most -- would be spent making do with whatever Social Security pays out and relying on the charity of family and/or strangers for the rest.

It's probably not a good idea to plan to live forever (and spend accordingly little in any given year) either, but planning to run out of money at your expected lifespan strikes me as clearly bad. For one, it is well known that wealthier people live longer. Maybe it's because of better access to healthcare, or because of less physically-stressful jobs, or because the very same traits that cause them to advance their career or business and save and invest also make them more likely to take care of themselves. Whatever the reason, the very fact that you are reading that Putnam paper (or, indeed, this blog) is a strong indicator that you are wealthier than average and are expected to live longer than average. But you won't find any adjustments in Putnam's paper for that since they use straight SSA actuarial tables and SSA is not concerned with your particular life expectancy, only with average life expectancy across your entire age group.

Bottom line: the study's assumptions about expected lifespans of investors are not appropriate for an individual investor. These assumptions result in unreasonably conservative allocations and unreasonably aggressive withdrawal rates aimed at maximum possible spending and running out of money at the end of expected lifespan. Individual investors are likely to live longer than the actuarial tables suggest and, more importantly, also have to plan for living longer than expected.

Annuitization to eliminate this longevity risk can be a very reasonable part of retirement planning. Unfortunately, annuitizing today will still get you nowhere close to the rates suggested by Putnam's paper. The reason for that is the second flawed assumption made by this study.

What Happens If Future Investment Returns Are Lower?

Putnam paper assumes historically average returns from cash, bonds, and stocks. These have been quite generous over the past century for US investors. Cash has historically more or less tracked inflation (in other words, 0% real return), bonds have delivered about 3% real return on average, and stocks delivered about 6% real return on average.

The problem is that we know with near absolute certainty that there is no way today's bond investments will deliver those historically average returns in the coming 10-20 years (the time span implied in the rest of Putnam's paper). The most optimistic possible outcome for bonds to deliver maybe 1% real return and personally I would count on negative 1% (-1%) to 0% real return instead. Cash is a little harder to predict but if today's rates stay with us for the next 10-20 years, then cash will be "earning" around negative 2% (-2%) annually after inflation. Stocks are, of course, the hardest to predict. They may yet end up delivering that historically average 6% real return but in my opinion it is much more prudent to plan on no more than 3-4% real instead -- and it could certainly end up even lower.

Unfortunately I don't have the time required to re-run Putnam's simulations to see whether and how these more realistic expected returns from cash, bonds, and stocks change "optimal" allocations (which are still subject to the very questionable life expectancy assumption we discussed earlier). What I can tell you with absolute certainty is that Putnam's recommended allocations coupled with their generous starting withdrawals  (e.g. 95% in cash and bonds coupled with 6+% withdrawals for 65-year old male) haven't a snowball's chance in hell to last even the expected lifetime!

This simple fact, by the way, is also why today's annuities won't pay you anywhere close to Putnam's numbers. Remember how earlier we said that Putnam's assumptions are appropriate for annuity or pension plans? Well, that was half of the story. The second half is that any such planning is based on returns available from different asset classes (primarily bonds) today, as opposed to those assets' historical averages. Annuity companies aren't going to assume 3% real returns from bonds just because that was the average of the past 100 years or so. They will look at what's available today and conclude that somewhere around 0% real return is what today's bonds might realistically deliver. Then they will price their benefits accordingly. A 65-year old male may very well be able to get a fixed annuity that pays out 6% today. The catch is that that payout will stay the same for the rest of his life, instead of being increased for inflation as this study assumes!

In Putnam's defense, it is very likely that their assumptions were more reasonable at the time of the actual study, in early 2011. Though not quite at historically average levels, bonds of two years ago yielded nearly 2% more than they do today at the long end of the yield curve (and I recommended locking into those rates several times on this blog). But cash yielded nothing back then, just as it does now, which means Putnam's cash-heavy portfolio suggestions were still highly questionable.

What Happens If You Hold More Equities?

The two big assumptions made in Putnam's paper make it next to useless for retirement planning for today's retirees. But the study does have a section that provides quite a bit of insight into retirement planning even today. What would have happened historically if instead of the extremely conservative cash/bond-heavy portfolios retirees held more equities? The answer is found in Appendix B of the paper.

What we find is that moving from an extremely conservative cash/bond portfolio to a much more balanced one (i.e. around 50% in equities) did not adversely impact portfolio's survivability almost at all, even given Putnam's generous (read: unrealistic) assumptions about returns from cash and bonds. For example, our 65-year old male withdrawing 6% of his portfolio in the first year and increasing withdrawals with inflation in subsequent years had the same very low (0.20%) chance of failure with full 75% of the portfolio in stocks and 25% in bonds as his did with Putnam's "optimal" portfolio of 5% stocks, 20% bonds, and 75% cash. Similar findings persist for other starting withdrawal rates, both for males and females.

This is a very important point and one that mirrors many other portfolio longevity/survivability studies I have seen: it simply does not make much difference whether one starts with a fairly aggressive portfolio (say, 75% stocks) or a fairly conservative one (say, 25% stocks). Historically both types of portfolios had about equal chance of surviving for reasonable lengths of time with reasonable rates of withdrawals. Sure, depending on the exact methodology used and exact assumptions, one study will conclude that 25% stocks is best while another will conclude that 75% stocks is best. The reality is that we have zero basis for putting much precision into our starting assumptions. All of these are guesstimates with very wide range of error. Quoting survival chances to fourth decimal place (e.g. 0.20% chance of failure) is completely meaningless. I'd say any and all such numbers should be assumed to have at least plus/minus 10% error ranges. That means that 90% estimated survival is as good as 100% survival and as good as 80% survival. Reading any more precision into it is fooling yourself.  Even the plus/minus 10% is probably too optimistic, but it's better than plus/minus 0.01% that Putnam study seems to imply.

Getting back to the most important point: although the "optimal" portfolios suggested by Putnam are very light on equities, the study's own numbers suggest that much higher equity allocations would have yielded statistically equivalent results. In today's environment where bonds and cash are essentially guaranteed to deliver significantly worse than historically average returns, my inclination would be to recommend an average retiree to hold more balanced portfolios. To me that means somewhere between 25% and 50% in equities (remember to not read too much precision into any of these numbers).

Obviously everyone's case is different and for some people it really does make sense to hold next to no equities. But those people are the ones whose portfolios absolutely dwarf their annual expenses.

If you are a 65-year old male retiring today, you have next to no chance of your portfolio's lasting as long as you do if you hold Putnam-advised 5% in equities, 20% in bonds and 75% in cash and start withdrawals at 6% of the portfolio and increase those withdrawals with inflation. Well, you do have a chance, but only if you die early -- not exactly the kind of chance you want to have. Of course, I would still guesstimate only so-so portfolio survivability even if you go with 50% stocks + 50% bonds (or 75% stocks + 25% bonds or 100% stocks) today with that 6% initial withdrawal but at least you'd have a chance.

Conclusion

The Putnam paper in question starts off with unreasonable and unrealistic assumptions for today's individual investments. That means that no matter how reasonable the paper's methodology is (and it does appear reasonable) its conclusions are deeply flawed and should be ignored.

The most interesting and relevant piece of information present in the paper is that very wide range of portfolio asset allocations result in essentially identical portfolio survivability. This is very much in line with other such studies. Because of this fact and because of the all-but-guaranteed subpar future returns from cash and bonds, I would urge today's retirees to have substantial allocation to equities in their portfolios (say, between 25 and 50% of the portfolio). I would also most certainly urge today's retirees to ignore Putnam paper's conclusions and to not let equity allocation drop to 5-10% levels that it advises.