Increasing Client Withdrawal Rates In Retirement

by: CFA Institute Contributors

By Isaac Presley, CFA

Evidence-based investing is about tipping the odds in our clients' favor. We can't predict the future, but we can give our clients the best chance at success.

So, if you subscribe to an evidence-based approach and work with retirees, pay attention to two recent studies that aim to tip the odds when it comes to spending from a portfolio.

For retirees, success means, in part, not outliving their money. The maximum amount they can spend from their portfolio is unknowable in advance, so a margin of safety must be built in. Hence the prevalence of safe withdrawal rate guidelines like the 4% rule.

But based on the year a client retires, the amount they could have spent when looking backward will vary widely based on the market's performance during their initial retirement years.

With a built-in margin of safety, if a retiree generates above-average returns but has no desire to leave a big estate, they run the risk of underspending. But most retirees will worry about below average returns and how they affect their chances of running out of money.

The two research studies focus on this last point and provide valuable recommendations to increase a portfolio's safe withdrawal rate and reduce the chances that clients may fall prey to an unlucky retirement date.

Let's dive in.

"Reducing Sequence Risk Using Trend Following and the CAPE Ratio"

In this CFA Institute Financial Analysts Journal® study, the authors consider a problem facing retirees: Withdrawal rates depend to a large degree on the year of retirement.

According to the research, on a portfolio invested in the S&P 500, a historical real 20-year perfect withdrawal rate, which depletes the initial balance at the end of 20 years, ran as low as 4% in 1930 to as high as 15% in 1949. Other studies have shown broad variation in safe withdrawal rates, even within a diversified portfolio.

Wade Pfau, CFA, shows that a 30-year inflation-adjusted sustainable withdrawal ratefor a portfolio invested 50% in stocks and 50% in bonds has ranged between around 4% for a retiree in the 1960s to a high of nearly 10% for the lucky retiree in the early 1980s.

In short, it's not just the level of returns that matters, the order, or sequence, of returns is also critical.

So, can this sequence risk - unlucky retirement date timing - be reduced by investing in a way that smooths the sequence of returns, or more specifically, reduces the magnitude of losses? The authors write:

"[A]lthough the order of returns cannot be predicted, it may be possible to produce investment strategies that offer substantially reduced return volatility or, more precisely, a much reduced drawdown of returns, because reduced volatility in itself is insufficient to secure a high PWA."

Their approach is to compare two portfolios: one that is 100% S&P 500 and the other S&P 500 with trend following. For the trend-following portfolio, they use a simple 10-month moving average that holds the S&P 500 if the current price is above the average over the last 10 months and holds cash if it is not.

"This all-equity portfolio may be considered rather unlikely as an investment choice in practice," they observe, "but it serves to illustrate our key points regarding the choice of investment strategy." They add that perhaps the portfolio shouldn't be so unusual: "A surprising result may well be that a 100% equity portfolio is not such a bad idea, providing that one overlays it with a trend-following filter."

The result of their comparison? A simple and transparent trend-following approach that "tends to smooth returns can offer superior perfect withdrawal rates across practically the whole range of return environments. It is this smoothing of returns that leads to a better decumulation experience for virtually all investing time frames."

They reach the finding in two ways: First, by looking at historical returns, in the order received, and then by using "Monte Carlo techniques to draw 20 years of returns, one at a time, with replacement."

Figure 1 depicts the perfect withdrawal rates over time and how they've varied between 4% and 15%. Additionally, in the worst-case scenarios for the S&P 500, the portfolio with trend following helps to improve the withdrawal rate.


Real PWR for 20-Year Deculumalation with 100% US Stock Investment: WIth and without Trend Following


In the Monte Carlo Analysis (Figure and Table 2), again the biggest improvement is on the low end by reducing the magnitude of drawdowns (a difference in maximum real drawdowns of 76.8% vs. 34.9% for the S&P 500 and the trend-following portfolio, respectively). Up to the 50% percentile, the trend-following portfolio shows a 1% or greater increase in withdrawal rates.


PWR as Percentage of Initial Balance: 20,000 Simulations of 20-Year Decumulation Using Annual S&P Real Returns with and without Trend Following, 1872-2014


Real PWR Percentiles as a Percentage of Initial Balance, 1872-2014

Source: All charts above courtesy of CFA Institute Financial Analysts Journal®.


Increasing Withdrawal Rates with Managed Futures

In the second study, Andrew Miller, CFA, highlights another problem. As he puts it, if forecasts about lower future returns given current yields prove true:

"[R]etirees are faced with a decision about either increasing the amount of assets they need to support their retirement withdrawals or slightly altering their portfolio composition, altering the composition is perhaps the more palatable of the two choices."

To alter the composition of a retiree's portfolio, Miller says that adding a modest allocation to managed futures to portfolios in the decumulation stage can increase the sustainable withdrawal rate.

The bulk of the paper looks at two different portfolios. One is composed of 50% stocks and 50% bonds. The other of 50% stocks, 40% bonds, and 10% managed futures. These portfolios are compared in three different scenarios - historical returns, a low returns scenario adjusted for current starting yields, and to further stress test the result, one that arbitrarily reduces the Sharpe ratio of managed futures allocation to 0.50, roughly cutting the return in half from historical results.

Miller finds that by adding a 10% allocation to managed futures, the withdrawal rate of the portfolio dramatically rises without increasing the portfolio's overall risk.

The withdrawal rate spikes by 20% (from 4.0% to 4.8%) when modeled with historical returns, by 23% (from 3.0% to 3.7%) assuming returns with current yield adjustments, and by nearly 17% (from 3.0% to 3.5%) even when assuming current yield adjustments and an arbitrarily reduced Sharpe ratio for managed futures.

Using the historical data, by adding managed futures to a $1-million portfolio, a retiree could sustainably spend an inflation-adjusted $48k annually vs. $40k without the allocation. Put another way, to maintain the $40k annual spending, a client could retire earlier with a portfolio balance of $833k rather than a $1-million portfolio under the 4% rule. Similar improvements, albeit with lower withdrawal rates, are observed when using yield-adjusted returns as well.


Safe 30-Year Withdrawal Rates

Safe 30-Year Withdrawal Rates

Source: "Improving Withdrawal Rates in a Low-Yield World," by Andrew Miller, CFA


So what does this research tell us about how we can better serve our clients?

We can't control the performance of the market when our client retires any more than we can the weather. But we don't have to leave their fate in the hands of Mr. Market.

A risk-management strategy may be just the ticket to increase their chances of success.

Disclaimer: Please note that the content of this site should not be construed as investment advice, nor do the opinions expressed necessarily reflect the views of CFA Institute.