For many years, investors have had two broad objectives for their portfolios—current income and future growth. Fortunately, fulfilling these objectives was a reasonably simple exercise, as investors have had the opportunity to select from two main types of assets to support these needs. These two asset classes—fixed income securities and equity securities—uniquely supported income and growth. Thus, a key component of portfolio construction was to properly determine this mix—a suitable bond portfolio to support cash flow needs and equities for future growth.
These days, it’s not so simple. As most investors well understand, bond yields are now as low as they’ve been in nearly 60 years. “Bond Yields,” this page, which shows highly rated corporate bond yields since 1919, illustrates the problem. The bond portion of a portfolio most likely can’t support an investor’s income needs.
Indeed, with interest rates at these historically low levels, investors are faced with critical new questions:
• What’s the role of bonds in a portfolio?
• How can a portfolio support income needs?
One common way to handle this dilemma is to view portfolios on what’s known as a “total return” basis. In other words, rather than split the portfolio into income and growth components, view the overall portfolio as supporting both needs. Thus, rather than using only the yield of a portfolio to support income, an investor can rely on both yield as well as capital gains. And, with equities having a higher expected return over time, this might be viewed as a rational solution.
In a total return framework, bonds are viewed as having a different role. Rather than being seen as the income component, fixed income is viewed as a risk reducer. As we all well understand—particularly in light of the events of the last few years—while equities have a higher expected return over longer timeframes, they can be very volatile over shorter periods. It’s worth explaining why this is important for portfolios with income needs. When investors make periodic withdrawals from portfolios, the path of returns comes into play. It’s not just important where the portfolio goes, it matters how it gets there. If we examine long-term return data (Ibbotson), equity returns have been approximately 10 percent per year. Stock returns have been far from a straight line. On average, the equity market falls about once every four years and the median decline has been roughly 15 percent. To solve for total return, this means that the return during the positive years has been approximately 20 percent. And, investors can’t plan for a negative return merely once every four years—they often come in clumps.
“Compounded Return,” this page, shows two examples with the same 12-year compounded return, each with nine positive years and three negative years. The difference is the path. In the “Early Rising Market,” the negative years come at the end of the period, and in “Early Falling Market,” they come in the beginning.
In looking at the compound return for the full 12-year period, both are the same—10 percent. But, the return path is very different, and the impact on portfolios supporting spending or periodic withdrawals is significant. Let’s consider a simple example of a portfolio with an initial value of $1,000. This portfolio supports initial spending of $50 per year with an inflation rate of 3 percent per year. Over a 12-year period, the spending will grow from $50 per year or 5 percent of the initial portfolio to $71.30 per year or a bit more than 7 percent of the initial value. With compound annual returns of 10 percent, this spending level, on the surface, doesn’t seem problematic. However, “How Spending Impacts Return Paths,” p. 39, plots the value of these two portfolios during their 12-year life, and the differences are stark.
In the “Early Rising Market” example, the initial $1,000 grows to nearly $2,300—a compound return of 8.7 percent. But, in the “Early Falling Market” example, due to withdrawals depleting the portfolio during the negative years, the final value is just shy of $1,250—a much lower compound return of 2.2 percent. With the early negative years, the final value is 46 percent less than the early positive years. What might be nearly as important is that the value of the portfolio impacted by the early negative returns falls to roughly half the initial value of $1,000 in Year 3. Thus, even though the portfolio successfully supported the spending during the full period, it’s quite possible that, after withstanding the initial decline in value, the investor might have been sufficiently worried to feel forced to make a shift in asset allocation at the worst possible time. So, while using a total return strategy might make sense, merely relying on long-term returns doesn’t appropriately deal with cash flow needs and client risk tolerance. While bonds may not have their historical ability to support cash flows, perhaps there may be a way for a fixed income allocation to reduce risk in a tangible way.
Risk Versus Consequences
One of the key problems with portfolio risk modeling is that it’s often been used to show what’s known as a “confidence band” for future portfolio values. Generally, investors are shown a chart describing some type of “best-case, worst-case” scenario, with the future described by a probabilistic model. I’ll leave a detailed description of this type of modeling for another article, but the drawback to this kind of analysis is that it often ignores the true worst-case scenario. For example, if a probabilistic model runs 10,000 trials of possible future portfolio behavior, the analysis will most likely identify the “worst case” using the 5th or perhaps the 10th percentile of the 10,000 trials. By using the 5th percentile, the model is stating that there’s a 95 percent chance that the portfolio value will be higher. Yet, there’s a 5 percent chance that it will be lower—how much lower we can’t determine. Nor do we know the consequences of this 5 percent chance of a worse outcome. There’s a very large difference between risk as defined by return volatility and the actual consequences that result from the event. Much of the financial community got a taste of this disconnect during the financial crisis of 2008, when previously unpredictable events became common occurrences. As Mark Twain reportedly said, “there are lies, damned lies, and statistics.”1
One way to use fixed income in a portfolio is to create a spending reserve. While bond yields may be too low to support spending purely from the interest income, it can be interesting to consider the bond portfolio as a wasting asset, built to support cash flow requirements for a set time. This kind of analysis assumes that the fixed income portion of the portfolio will support spending until the asset is depleted. While, in reality, it’s unlikely that a portfolio would be managed this way, this analysis creates a certain sense of security and defined risk tolerance for the investor. The following hypothetical is indicative of this process.
Let’s revisit our example of a $1,000 portfolio supporting initial spending of $50 and growing at 3 percent inflation. Rather than focus on specific return paths or on a probabilistic model, let’s use the inputs in “Create a Spending Reserve,” p. 40.
Using this analysis, we have short-term assets made up of low return, low risk fixed income securities and long-term assets with higher return and higher risk holdings. Further, we’re assuming that we have
30 percent of our assets initially held in our reserve portfolio to support spending and that this portfolio will support all spending until it’s depleted.
This type of analysis identifies a period of relative certainty regarding spending. While long-term assets may indeed be volatile, if an investor doesn’t need to draw upon the long-term assets, this eliminates the issue of path dependency. And, to the extent that an investor recognizes that he has a specific number of years of spending essentially locked up, the ability to withstand market volatility without emotional or reactive decisions expands. “Total Portfolio Analysis,” this page, describes the result of this analysis.
It shows the value of the short-term portfolio represented by the purple bars along the horizontal axis. In this analysis, the short-term assets are able to support portfolio spending for approximately six years. The median expected value of the portfolio in 20 years is $869 and in the 5th percentile, the portfolio is depleted in Year 19. But, the critical information is the term over which the short-term assets can support cash flows. Is six years enough? For some clients, the answer may be yes. Others may require more or, perhaps, less relative certainty. But, this kind of analysis allows a fixed income portion of an investment portfolio to reduce risk in a tangible manner.
Additional Credit Risk
As I’ve stated, in this current interest rate environment, U.S. Treasury securities and highly rated bonds have very low yields. Not surprisingly, many investors have attempted to boost yields by either taking on additional credit risk or by extending maturities and taking additional interest rate risk.
In any analysis using short-term assets, it’s crucial that this type of asset be truly low risk. Taking on additional credit exposure is something more appropriate for growth portfolios. And, extending duration (maturity) increases the risk from potential inflation and rising rates in the future. Adding any kind of increased return volatility to reserve assets defeats the point of the strategy.
1. Mark Twain, Chapters from My Autobiography, North American Review (1906).