Some of the most common precepts of retirement planning don’t hold up under close scrutiny. Here are seven misconceptions, and how they could hurt your clients.
1. Betting on Asset Allocation
Claim: “Asset allocation contributes over 90% to a portfolio’s success”
The first serious look at the importance of asset allocation was the study by Gary P. Brinson, Randolph L. Hood, and Gilbert L. Beebower in 1984. The authors analyzed data from 91 corporate pension plans with assets of at least $100 million, over a 10-year period beginning in 1974. In the literature, this is generally referred to as the “Brinson Study.” Subsequently, the analysis was expanded to include an additional ten years of data and “Determinants of Portfolio Performance II” was published in the Financial Analysts Journal, January/February 1995.
Their conclusion was that the components of the difference in success of a portfolio are:
- Asset allocation: 93.6%
- Security selection 2.5%
- Other: 2.2%
- Market timing: 1.7%
Here is the problem: The findings of the Brinson study cannot be applied to individual retirement portfolios for the following reasons:
- The dynamics of cash flow in a pension fund are entirely different from those in an individual retirement account. A pension fund has a continuous inflow of money over time. In an individual retirement account, inflows of money occur mainly during working years. After retirement, there are usually no more inflows, only outflows. Mathematically, a pension fund is an “open-perpetual” system; an individual retirement account is a “closed-finite” system.
- When and if there is a shortfall in a pension fund, contributions can be increased to meet it. On the other hand, with an individual retirement savings, there is usually no such opportunity.
- In an individual retirement account, once withdrawals start, the adverse effect of “reverse dollar cost averaging” becomes important. In a pension fund, since there is a continuous inflow of money concurrently, this is insignificant.
- In an individual retirement account, inflation is important. Withdrawals must be increased over time to maintain the same purchasing power for the retiree. In pension funds, there is no such concern; as inflation goes up, salaries and pension contributions increase. Also, many pension funds have limits and constraints on how the retirement payments are indexed, if at all. Individual retirees holding their own saving accounts do not have that choice; their expenses must be met.
- The Brinson study does not include two of the most important components of the success or failure of distribution portfolios: withdrawal rate and sequence of returns. If you do not include the most important factors in the analysis, then less important factors – such as asset allocation – become most important.
- Another factor is portfolio management costs. For an individual retirement portfolio, it is significantly larger than typical pension funds.
- The timespan of the Brinson study is twenty years. This is not only too short, but it covers a single secular bullish trend, arguably the “luckiest” 20-year time period over the entire twentieth century. The study’s data misses significant adverse events that happened during secular bear or secular sideways market trends.
Figure 1: Time period covered in the Brinson study
Claim: “Lowering a portfolio’s volatility increases its life”
Traditionally, asset allocation manages portfolio volatility within a client’s risk tolerance, which is the vertical axis in Figure 2. However, asset allocation has little or no effect on a portfolio’s life, reflected in the horizontal axis.
Figure 2: Components of the portfolio value versus time
There are three factors that affect the horizontal axis (i.e. a portfolio’s longevity):
- withdrawal rate;
- sequence of returns; and
Until all of these factors are dealt with, asset allocation does not help longevity much.
Take a 65-year-old investor, retiring this year. He wants to plan until age 95. His retirement savings are valued at $1 million. He needs to withdraw $60,000 each year, indexed to actual inflation. On the equity side, he expects a 2% average dividend yield and pays 0.5% management fees.
Let’s look at how his portfolio would have performed if he were to start his retirement in any of the years between 1900 and 2000. We aftcast six different asset mixes:
Table 1: The impact of asset allocation for various withdrawal rates
(Equity / Fixed Income)
|Probability of Depletion by Age 95||Median Portfolio depleted at Age|
|80 / 20||67%||87|
|60 / 40||74%||87|
|40 / 60||78%||86|
|20 / 80||91%||86|
|100% Fixed Income||95%||87|
This table shows that regardless of asset allocation, the median portfolio depleted at ages 86 or 87. If you want asset allocation to be an effective tool, find ways of reducing the withdrawal rate to below sustainable (see the course, Purpose-Driven Sustainable Withdrawal Rate): delay retirement, save more, spend less, find a part-time job, downsize, rent basement, etc.
The truth about asset allocation: it’s is an effective tool to manage volatility. However, if the withdrawal rate is larger than 4%, don’t bet on asset allocation for converting a red-zone client into a green-zone client. (see the course, Lifelong Retirement Income – The Zone Strategy).
2. Equities and inflation
Claim: “Over the long term, equities always beat Inflation.”
Sometimes they do, sometimes they don’t. It depends on what the long-term (secular) trend is and which equity index we are looking at.
Secular Bullish Trends: During the last century, there were three long-term bullish trends: 1921-29, 1949-66, 1982-2000, covering 43% of time. During these periods, inflation was low and equity markets overcame inflation readily.
Because Canadian markets are more commodity oriented, they underperformed U.S. markets during these time periods, as depicted in Figure 3.
Figure 3: Inflation, DJIA and TSX in secular bullish trends (1921-1929, 1949-1966, 1982-2000)
Secular Bearish Trends: This is when inflation is either very low or negative (deflation). However, losses in a bear market overwhelm any positive effect of deflation (i.e. reduced withdrawals). During the deflationary time period between 1929 and 1943, neither Canadian nor U.S. equity markets were able to beat inflation.
Secular Sideways Trends: There were two secular sideways trends last century: 1900-21 and 1966-82. Generally, inflation peaked towards the end of these trends, forcing the retiree to withdraw more from her portfolio for the rest of her life. During these time periods U.S. indices lagged inflation miserably.
Figure 4 depicts inflation, the DJIA and the TSX. Note that historical data for the TSX index does not start until 1919, therefore, there is no graph for TSX for the 1900-20 secular sideways trend.
Figure 4: Inflation, DJIA and TSX in secular bearish and sideways trends (1900-1921, 1929-1949, 1966-1982)
During the 1966-82 period, Canadian and U.S. markets behaved differently. Since a large component of the Canadian equity market includes commodity type companies, they outperformed U.S. markets during inflationary time periods. Even though neither index was able to beat inflation, Canadian equities did much better. U.S. equities experienced a double whammy of bad sequence of returns and high inflation, whereas the Canadian equity index experienced a much milder bad sequence of returns and relatively good inflation protection.
Make sure you are not holding excessive amounts of U.S. and other foreign equities in retirement portfolios. Many in the financial industry argue that since Canada is only 3% of the world economy, therefore you need to allocate a large chunk of assets to foreign stocks. In this case, size is not as important as quality. Financial statements and disclosures of Canadian companies are likely more reliable than elsewhere. Also, most Canadian companies have significant amounts of international business, which is your de-facto international diversification.
The truth about equities and inflation: Equities do not always beat inflation, but Canadian equities do better than others. No more than 10% to 15% of the portfolio should be allocated to foreign equities. Also consider allocating up to 10% to real return bonds.
3. Making Assumptions on Average Growth, Inflation and Life Expectancy
Claim: “Assume a 5% annual growth rate and 3% inflation over the next 20 years”
The main difference between a Gaussian and a non-Gaussian mindset is this: the Gaussian mindset uses averages in calculations. The non-Gaussian mindset uses extreme outcomes and avoids averages in decision-making.
In the financial media, you will invariably see articles that say things like this: “If his average portfolio grows 5% annually and the inflation is 3% per year, assuming he lives until age 85 (his life expectancy), he should be able to withdraw $30,000/year from his $500,000 portfolio….”
Here is how this assumption manifests itself on the chart:
Figure 5: Forecast using average 5% growth rate, 3% inflation, 6% initial withdrawal rate
Here are flaws of using average assumptions:
Average Life Expectancy: This is the age at which half the cohort will be dead and the other half survives. If you design for the average life expectancy, you are in effect planning only for half of your clients.
A better way of managing longevity risk is using an age beyond which the survival rate is 10%. That means using 95 as the age of death for male clients and age 97 for female clients or couples as a minimum.
Average Portfolio Growth: In real life, there is no such a thing as average growth rate; it only exists in statistics:
- One bad sequence of returns can turn an otherwise great track record to a dismal one.
- Studies on investor behavior show consistently that an average investor lags the average market performance by a factor of several percent.
- It is also well documented that most funds do not beat the market index.
- During retirement, withdrawals cause a reverse-dollar-cost averaging effect, which can reduce annual returns by up to 2%.
A better way of managing this risk is to ignore average growth rates totally. Instead, make sure that the probability of depletion does not exceed 10% at the age of death using actual historical numbers, i.e. the aftcast.
Average Inflation: One of the largest risk factors during retirement is inflation. Here, just as with the growth rate, the average is meaningless. Historically, inflation was higher than the average slightly over 50% of the time. Using the average exposes the retiree to a significant inflation risk. Plan in such a way that the loss of purchasing power does not exceed 10% for essential expenses throughout retirement.
Figure 6 depicts the aftcast and the forecast with unsustainable withdrawals based on assumptions using averages. The aftcast shows that these withdrawals are definitely unsustainable even if the age of death was 85.
Figure 6: The aftcast and the forecast, unsustainable withdrawals
When we change the age of death to 95 and keep the probability of depletion to under 10%, as we should, then the sustainable withdrawal becomes $18,000/year, 40% lower than the $30,000 calculated earlier using “averages” (Figure 7).
Figure 7: The aftcast with sustainable withdrawals
The truth about using the averages: They apply only to large groups and they should never be used for individuals.
4. Frequent Rebalancing
Claim: “Frequent rebalancing reduces volatility”
When it comes to rebalancing, most investment professionals believe “often is better.” Rebalancing is done supposedly to reduce volatility. But does it? How does it affect portfolio longevity? Let’s answer these questions.
Volatility has two components. The first component is short-term, random fluctuations, generally within a one-year time horizon. Every second, every minute, every day, some event happens somewhere in the world that influences investor psychology. As investors make trading decisions, markets move up or down. This is how random volatility is created.
The second component of volatility occurs over the longer term. Markets respond to the collective expectations of investors and a trend forms. As these expectations become more and more popular among investors, security prices move further with that trend. Gradually, extremes are created and they are invariably followed by a trend in the opposite direction.
If you consider one year as short term, then rebalancing annually (or more often) cannot reduce volatility. Rebalancing can reduce volatility only after an observable trend. An observable trend can occur for two reasons:
- After a cyclical market trend, which typically lasts four years; or
- If withdrawals are high, that in itself creates a downtrend in the portfolio value.
When does a portfolio experience an observable trend? There are several known market cycles: the 54-year Kondratieff cycle, 10-year decennial cycle, and the 4-year U.S. presidential election cycle, to name a few. We focus on the U.S. presidential election cycle as it is the shortest market cycle that is meaningful to retirement planning.
Let’s look at an example: Steve, 65, just retired. He has put aside $1 million for his retirement, with $400,000 in equities and $600,000 in fixed income. He needs $50,000 in income each year, indexed to inflation.
Figure 8: Retiring at the start of the 1921 secular bull market
Figure 8 shows the portfolio value if Steve were to retire in 1921, the beginning of the first secular bull market of the last century. The green line shows the portfolio value if rebalanced every four years at the end of the U.S. presidential election year. The red line shows the portfolio value if rebalanced annually. At the end of 30 years, Steve was one million dollars richer if he rebalanced every four years at the end of the U.S. Presidential election year than if he rebalanced annually. The volatility was about the same for both. Imagine that: your client makes more money, and you make more trailer fees for less (rebalancing) work!
Figure 9 shows Steve’s portfolio value if he were to retire at the beginning of 1929, the beginning of a secular bear market. At the market bottom of 1932, Steve’s portfolio experienced a smaller loss when rebalanced every four years than if he rebalanced every year. The portfolio that was rebalanced every four years provided Steve with 30 years of income. On the other hand, if rebalanced annually, the portfolio would run out of money after 21.5 years. Rebalancing every four years on the presidential election year increased the portfolio life by a respectable 40%.
Figure 9: Retiring at the start of the 1929 secular bear market
Similar analysis for all years between 1900 and 1999 demonstrates that:
- Portfolio volatility was about the same whether you rebalanced annually or once every four years on the presidential election year.
- In secular bull markets, rebalancing too often generally stunted portfolio growth.
- In secular bear markets, rebalancing too often compounded losses. Synchronizing the rebalancing activity with the U.S. presidential election cycle preserved one’s capital during black swan events more effectively.
- In secular sideways markets, portfolio longevity was not correlated to the frequency of rebalancing.
The truth about rebalancing: Rebalancing too often can hurt portfolio longevity. If the withdrawal rate is below 4%, then rebalancing once every four years (at the end of U.S. presidential election year) can provide a better outcome in retirement portfolios.