The Dreaded Sequence of Returns


The concept of sequence of returns has come up twice in two different spots in the last couple of days. One was this article by Gil Reich (read the comments too) as well as in the comments of my Facebook feed when I posted a link to a recent article I published on TheMaven.

Sequence of returns is simply the string of returns you get (or the market provides) over some period of time. Last year, maybe you were up 13%, this year maybe up 18% and next year maybe up 2%. That would be an example of a short sequence of returns. Longer term, the average annual return of the market is a long term sequence and that number (differs depending on the time studied) takes in all the great years, the terrible years and everything in between.

The retirement planning implication has to do with when you retire and what the market then does. Someone who retired on Dec 31, 2006 and started drawing from their portfolio was likely adversely affected by the sequence of their returns. In 2007 the S&P 500 was up a little and then of course the in 2008 it dropped by close to 40%. A 60/40 portfolio worth $500,000 on retirement day, might have been about even one year later assuming a $20,000 (4%) withdrawal. A year later that portfolio might have been worth $280,000 assuming a 40% hit to the equity portion of the portfolio, maybe a little better than that as bonds did well that year. A $20,000 withdrawal becomes a little more uncomfortable 6.6% and so if the portfolio was at $300,000 after the decline in 2007 it would be $280,000 after the withdrawal.

In 2009 the market of course bounced nicely but it took several more years to get back to its highwater mark (this happened in 2013) but if the retiree above kept taking withdrawals then it took quite a bit longer to get there. I did a little crude math whereby the equity portion of the portfolio tracked the market exactly and half of the withdrawal, or $10,000, came from the equity side. Starting from $300,000, tracking the market down and then back up with withdrawals along the way I get the equity portion sitting today at $421,000.

Doing something similar with the $200,000 in fixed income using the iShares Aggregate Bond ETF (AGG) as a proxy and accounting for $10,000 of the annual withdrawal gets the bond portion down to $105,000. Not surprisingly the bonds could not keep up with the withdrawals and the total account value then stands at $526,000, this is arguably a lost decade. It could have been worse though, they in part were bailed out by an amazing bull market run. A bear market two years ago, still would have been a very long bull market, might have been a death blow for this retirement plan. The next bear market could do serious damage still, if equities come back slower. This scenario as also been bailed out by mostly benign inflation. Higher inflation could have been a problem.

I concede any flaws in the process, it is merely an illustration and how things could have played out.

Conversely had this person retired on December 31, 2010 with the same $500,000 the equity portion of their portfolio would now be $551,000 (from a starting point of $300,000) using the same methodology versus $421,000 (equity only) starting retirement four years earlier.

Sequence matters and of course a big part of it is luck. So how do you mitigate this threat? In Gil's above linked post he suggested selling everything once you hit your number and then doing some sort of slow dollar cost averaging to get back in. This can work especially if doing this comes fairly close to a top but that would be a matter of luck. The luck would be bad if the market goes up another 20% in 2018.

I believe a better way to go would be to set aside something like two or three years worth of expected cash need and keep the rest of the portfolio invested per what ever strategy had been employed all along. I think there could be some added nuance in terms of maybe a little more cash now, almost nine years into a bull market but maybe a little less cash two years into a bull market.

The difference in the two examples between $421,000 and $551,000 is huge and remember that is just equity, the fixed income would add another $140,000 approximately. Facing the unlucky side of that outcome, things would have been much improved if the retiree could have figured a way to not take his withdrawals. Is that a realistic solution? That might depend on how well someone has planned but someone thinking of retiring in the next year or two should explore mitigating this problem as there is a good chance, based on how long cycles tend to last, that they are vulnerable to an unlucky sequence.