Tuesday 20 May 2014

Retirement Income Design: Making Sure Your Portfolio Survives as Long as You

In last week's post, we briefly touched on some ideas - diversification and spending buckets - for minimizing the effects of poor returns early on during retirement, termed sequence of returns risk. Today we'll delve further into diversification and a number of other techniques for controlling sequence of returns risk and the follow-on risk of the portfolio running out of money during retirement.

1) Diversification
As we noted lest week, diversification works well through the counter-balancing effects of un- or weakly-correlated asset classes. Here are some ways to exploit this best:

  • Additional asset classes - Every additional asset that is less than perfectly correlated with the others helps, though the effect tails off. The basic combination to create the simplest possible portfolio of two-holdings is Stocks/Equities and Bonds. Additional holdings that help diversify include: real estate (REITs), real return bonds, cash/T-Bills, commodities or just gold, and foreign equities, sub-divided into developed markets and emerging markets.
  • Consistently low-correlation combinations of asset classes - One reason gold has been so useful in the Permanent Portfolio, is its consistent lack of correlation with every other asset class, including during the 2008 financial crisis stock meltdown when gold went up strongly. In contrast, in 2008 some asset classes, like all the equities and REITs, that had formerly exhibited beneficial reasonably low correlations suddenly became highly correlated. 
2) Low Volatility Funds
Within equity ETFs, in the last few years some funds have come on the market explicitly designed to be much less volatile than traditional broad cap-weighted ETFs while still being fairly representative of the broad market. Examples are the BMO Low Volatility Canadian Equity ETF (TSX symbol ZLB) and the Powershares S&P 500 Low Volatility Portfolio (NYSE: SPLV), both of which have, in the two or three years since their inception at least, been much less volatile (2 to 3% less annualized standard deviation of daily returns) than respective broad index funds, as can be seen by plugging their symbols (use ZLB.TO) into the handy InvestSpy.com calculator. In addition to ZLB and SPLV, there are a number of other domestic and foreign low volatility funds - see our review of the concept and the pros/cons in this post of January 2012.

Similarly, many dividend stocks and dividend ETFs exhibit less volatility than broad cap-weight index funds, as we saw back in January when we compared Canadian dividend ETFs.

Lower volatility can also be found among bond funds less exposed to interest rate shifts due to the shorter term and duration of their holdings, though there is the downside that returns will be less too. This can be seen, for example, in iShares Canada's table of fixed income funds.

A variation on the theme of reducing volatility is to directly balance the volatility risk of the portfolio holdings, as we showed in this post. The result, however, is a significant shift in weighting towards generally lower return bonds.

3) Rising Equity Allocation during Retirement
Completely opposite to what many people have long been told, as expressed in the rule of thumb to hold your age in more stable bonds and the rest in equities (e.g. a 65 year-old would have 65% in bonds and 35% in equities, a 75 year-old 75% in bonds and 25% in equities), retirement researchers Wade Pfau and Michael Kitces tell us in Reducing Retirement Risk with a Rising Equity Glide-Path that a retiree has less chance of running out of money at all before dying, and less of a shortfall (running out fewer years sooner) when a portfolio did run out early, if the equity allocation at age of retirement was a relatively low 20% to 40% and then was increased steadily year by year to the 40% to 80% range. This makes intuitive sense - having less of volatile equities when one is most vulnerable to sequence of returns risk early limits the possible downside. 

The study looks at a number of scenarios, a sobering one being future returns much lower than past history: 3.1% (vs 6.46% historical long term average) compounded after-inflation for equities and more or less nothing 0.06% (vs 2.35%) for bonds. Using those lower assumptions, which reflect current market conditions, the initial 4.0% withdrawal rate of portfolio value at start of retirement, upped yearly by inflation thereafter, cannot be sustained. Only a rate of 3.0% withdrawal would be feasible, per the chart below in the outlined green cell, and even that leaves a 10% chance of portfolio failure before death assumed to be 30 years after retirement. Scary! 

Of course, a retiree is not obliged to continue taking out 4% annually no matter what. Probably that wouldn't happen if a retiree saw money running out, spending would be curtailed. There are other portfolio withdrawal strategies that vary the amount of money withdrawn each year.

4) Flexible and Variable Portfolio Withdrawal Strategies
  • Constant percentage of portfolio value at start of year - A retiree could take a snapshot of portfolio value every January 1st and withdraw 4% of the balance, or some other percentage. The problems with this approach are that i) market shifts could mean large differences in amount available to spend, as even those with diversified portfolios have experienced and ii) setting the percent withdrawal too high, much above 5% or so, would still risk depleting the portfolio. Though by definition the portfolio would never run out of money, it could get very small quickly.
  • Floor and ceiling spending - This type of strategy sets a minimum real dollar spending level, after years of down markets and a maximum for good market return years. Financial advisor and author William Bengen, in his classic book Conserving Client Portfolios During Retirement, calculated that based on historical returns in the USA, which slightly exceeds those in Canada, a spending range of +/- 5% permitted a starting withdrawal rate of 4.9%. On a $500,000 portfolio at start of retirement, that would allow withdrawals of $24,500 in the middle and a range from $23,275 to $25,725.
Fund provider Vanguard makes an insightful comparison of the two above strategies plus the fixed 4% case in this paper

Another Kitces article that seems not to be available free online discusses a series of variables, such as diversification through additional asset classes, fees, differential taxes on interest, dividends and capital gains, spending flexibility, varying time horizon, good/bad economic environment, and their impact on possible withdrawal rates. AUM in a Box blog summarizes the key findings in this post.

One approach that no one seems to propose to retirees is to plan spending according to the rising-with-age percentage withdrawal rate that is imposed for RRIFs and other types of registered retirement plans in Canada. It may be a way for the government to force the deferred taxes in such plans to be paid back but it isn't a useful retirement income strategy, though ironically it does reduce sequence of returns risk by keeping withdrawals lower at the start of retirement.

5) Annuitize 
Purchasing an annuity, which pays out a pre-determined amount,  completely avoids the risk of market fluctuations in a retirement portfolio for whatever proportion or amount that is annuitized. The reluctance of the majority of retirees to buy annuities, which is what economists think is the rational action, puzzles economists to the point they call the reluctance the "annuity puzzle". The annuity puzzle is much debated.,as can be seen in advisor Michael Kitces' Annuities vs Safe Withdrawal Rates: Comparing Floor/Upside Approaches, which has many thought-provoking comments from other advisors who put forward practical reasons for and against annuities.

6) Safe Savings Rate
A final method, which applies to younger people near the start of their working and savings phase, is Wade Pfau's proposed Safe Savings Rate. According to Pfau a person can, based on historical data, set and carry out throughout his/her working life a rate of savings as a percentage of income that ensures a certain level of retirement income. 

7) Toss that Bucket
By the way, the possibility of using a strategy of different buckets was also examined by financial advisor Jim Otar of Retirement Optimizer. In his 10 Bucket Strategies that Don't Work, he calculated what would have happened using actual market returns for any retiree from 1900 onwards and found none of the ten bucket strategies provided failsafe portfolio survival.

Pfau's blog post includes a note investors must remember - despite all the on-going research on retirement, it is still very hard to assure success and to know if we are a path that will ensure success: "The new article also emphasizes how hard it really is to know if one is on track to meeting a wealth accumulation target by a given date." Even the experts cannot tell, so we individuals must stay flexible and be ready to adjust as required.

Disclaimer: This post is my opinion only and should not be construed as investment advice. Readers should be aware that the above comparisons are not an investment recommendation. They rest on other sources, whose accuracy is not guaranteed and the article may not interpret such results correctly. Do your homework before making any decisions and consider consulting a professional advisor.

No comments: