Wednesday, November 5, 2014

RIIA Webinar -- Sequence of Returns Risk

Powerpoint slides for the Retirement Income Industry Association webinar that I presented November 5, 2014 on Sequence of Returns Risk can be downloaded here. RIIA will soon post the video of the presentation and I will provide a link to it from this page as soon as it is available.

Whether you attended the webinar or viewed it later, please post any questions in the comments section below.

Following are responses to questions that were sent to RIIA.

Q: But the risk to a saver in the sense of IMPACT to portfolio is higher nearer retirement because the value of the portfolio is typically at highest point

A: True. As I mentioned in the webinar, a typical retiree will make the largest bets at the end of saving and the beginning of spending. That’s why it’s important to lower your equity allocation during those times to place smaller bets on stocks. Sorry if that wasn’t clear. I tried to circle both areas of the graph with my cursor.

Q: Aren't portfolios in accumulation less susceptible to the order of returns as long as you rebalance into relatively underperforming assets over time...and are passive as well?

Good question! Rebalancing seems to help, at least based on historical data, in both saving and spending stages. Using the RetireEarly Homepage model, rebalancing improved a 91% failure rate to 95% with 4.7% spending in the past. But, most SWR studies rebalance, including Bengen’s original work, so the spending rates you see are probably based on rebalanced portfolios already. Not rebalancing would probably make sustainable spending rates a little smaller.

Q: Talk about bond ladders of zeros as a strategy in retirement please

A: TIPS Bond Ladders are an attempt to build a better personal annuity in a safer way than SWR, which attempts it with stocks. Treasury ladders have no credit default risk, while annuities are subject to the financial strength of the insurance company. Annuities have no residual value after you and your spouse die (there are some protection features available at a cost), but your heirs can receive any unspent TIPS bonds left in the ladder. You can’t really build your own annuity either way, since annuities pool longevity risk. You can build a 30-year TIPS Bond ladder if you can afford it, but if you live 35 years, an annuity will still provide income.

If you search my blog at The Retirement Café for TIPS Bond Ladders, you will find several posts with more information (here, for example.). I plan to write on the topic again soon, so please check back.

I would say that the major problem with TIPS Bond ladders today is their high cost, given historically low interest rates (the same is true of annuities). I’d recommend a 15-year rolling TIPS Bond ladder.

Q: Why is credit/default risk not considered in bond ladders? It was mentioned home loans from the debtor side.

A: It is considered in bond ladders if the bonds aren’t Treasuries. The U.S. government is prohibited by law from defaulting on a Treasury bond. I only recommend Treasury bonds for retirement ladders.

Q: How does this all relate to a risk-adjusted return approach to portfolio allocation? As you noted, different asset types have different returns (and even within bonds)?

A: The chart I showed from Bengen shows the effect of equity allocation on SOR Risk. It doesn’t matter how you arrive at a particular equity allocation. However you arrive at an equity allocation, it appears that you will increase SOR Risk with greater than about 70% equities.

A:  Yes. Spending directly impacts SOR Risk and taxes are more spending.

The impact of inflation on your portfolio depends on the assets you hold. TIPS and Social Security benefits are adjusted for inflation, so there is little impact. Stocks perform poorly when inflation is high, but their returns typically make up for it over the long term. Payments from a pension with no COLA adjustments would be worth less and less over time, requiring you to spend more from your risky portfolio and increasing SOR Risk by increasing spending.

Q: Any comment on Kitces analysis of how SOR risk and inflation periods have played out in history?

A: Only these. Inflation isn’t a problem for the foreseeable future. And future market returns may not look like past returns. There are more good arguments that growth will be slower than there are that it will increase. I wouldn’t bet my standard of living that things will turn out in the future the way they have in the past.

Q: Great presentation!  Followed you right to the end.  
A: My favorite question!

Q: It seems like sequence of returns risk is magnified in a retirement where on uses a constant dollar spending approach (Bengen).  Couldn't one solve for sequence of returns by being more flexible with their spending, such as using a constant percentage approach (theoretically never running out of money though spending could potentially decline to a low number at some point)?

A: Yes. I wrote a series at The Retirement Café entitled “Clarifying Sequence of Returns Risk” showing that a constant percentage approach is safer. But as William Sharpe says, “Isn’t it self-evident that your spending should depend on how much money you now have?”

I’ve been retired for 10 years and how much money I used to have when I first retired doesn’t seem to matter to anyone.

Q: Could one lower their sequence of returns risk by beginning retirement with a relatively low SWR level (say, 2.5%) and then slowly increasing the SWR up to the more traditional rate (about 4%) over the course of a few years at the beginning of retirement (say the first 5 years)?

A: Yes. Anything that lowers your spending from a risky portfolio lowers your SOR Risk. Some people can spend less early because they have a part-time job, for example.

Q: GREAT JOB - thanks Dirk I really like your blog!

A: We have a tie for best question!

Q: Professor Stephen Sacks and I have done some work on mitigating the effect of SOR (J. Fin'l Plng Feb 2012) using reverse mortgage credit lines.  Also, Professor John Salter and others have done similar work.  Could you comment on that? Thanks, Barry H. Sacks

A: Without commenting on reverse mortgages, because I have little exposure to them, what I do know is that they would behave much like an annuity, reduce spending from a risky portfolio, and thereby reduce sequence of returns risk.


  1. Do you think there are also sequence-of-expense risks in retirement? For instance, if inflation averages 3% over the retirement window, how much difference does it make if below-average inflation occurs earlier in the window versus later (assuming that not all income is inflation-adjusted nor needs to be)?

    And how about for health-care costs that typically rise faster than the CPI? Does sequence of expenses matter here? (I'm thinking partly about Medicare part B costs that have been level this year and will again be in 2015 but historically increase faster than the CPI. I'm also thinking that bumping up taxable income withdrawals at a faster rate year-over-year than the CPI can outpace annual CPI adjustments to tax rates and brackets, raising a retiree's effective federal tax rate.)

  2. Good question!

    There is "sequence of expense" risk in retirement, or said differently, sequence of returns risk includes expenses. Early portfolio failure isn't directly a result of low returns early in retirement. The direct cause is spending more than your portfolio earns. You can see this because a lower spending rate survives more periods of poor returns than a higher one, and because your portfolio can survive any historical period of poor returns if you lower the withdrawal rate enough.

    So, higher expenses early in retirement for any reason will drain more from your savings and increase the probability of portfolio failure and will have a greater impact than more spending late in retirement.

    Inflation, however, is a separate issue and depends on the sources of your retirement income. Social Security benefits have a COLA adjustment. TIPS bonds have an inflation adjustment. You can purchase annuities with a COLA adjustment, though they cost more (I think they're worth it). Stocks perform poorly when inflation is high, but tend to make up for it later with higher returns. To the extent that your retirement income comes from these sources, inflation won't make a lot of difference.

    Some pensions don't have inflation protection. Several senior expenses see more inflation than you can protect. To the extent you have these, prices will increase over time, but will have less impact on SOR risk because they are greater later.

    Thanks for writing!

  3. I liked your SOR slide presentation. What do you think of a rolling 10yr CD bond ladder. I am not very familiar with TIPS but I do like CDs. Will a CD bond ladder mitigate SOR risk? Also, holding individual bonds till maturity, does that mitigate SOR risk?

    1. Excellent questions.

      A rolling 10-year CD ladder would have no SOR risk, but then neither would individual CD's. CD's are not volatile. Their value is independent of stock market returns. There is also no interest rate risk, because you lock in the interest rate when you buy the CD. Therefore, there is no sequence risk.

      A CD ladder is a very safe way to insure that you will have a predictable amount of income to spend in each of those 10 years. Why would I prefer a TIPS ladder? TIPS bonds are arguably safer, but not much, and TIPS have inflation protection while CD's don't. But I agree that CD's are easier to buy than TIPS bonds.

      TIPS are individual bonds, so I'm guessing you are asking about a ladder of corporate bonds. That would also eliminate SOR risk, but would introduce credit default risk. The U.S. government can't default on bond payments by federal law, while a corporate bond issuer could. Like CD's, corporate bonds are subject to inflation, so I would prefer TIPS. Corporates pay higher rates of interest than Treasuries, but that's because they're riskier.

      The next risk of concern would be that buying a ladder of anything today is locking in historically low interest rates. A TIPS bond fund might perform better than a ladder over the next decade, but it also might not. It seems like the odds of rates going up outweigh the odds of them going farther down at this point, favoring funds over ladders. But, only with a ladder do you know exactly what you will get back and when.

      So, my preference for safety and inflation protection would be a TIPS bond ladder. A CD ladder would be fine if you're OK with no inflation protection. Corporate bonds aren't a good idea, even laddered. They have neither inflation protection nor protection from default.

      If you are in a high tax bracket (most retirees aren't), then laddered municipal bonds might make sense.

      Thanks for writing!