It’s common for financial advisers and planners to focus on expected returns of financial assets in helping clients reach their goals. With a targeted return in mind, they can construct a portfolio allocation that’s intended to lower overall risk, primarily by seeking to find a suitable combination of “uncorrelated” assets.
This approach is well and good, theoretically. However, as was demonstrated in spades in the 2008 meltdown, almost all asset classes became highly correlated, exposing clients to much greater portfolio losses than they or their advisers anticipated.
Risk budgeting involves identifying a client’s risk tolerance first, then building a portfolio that will achieve maximum expected returns without violating that budget.
First, let’s look at how we evaluate risk. Quantifying the many ways risk can be manifested in investment portfolios is no simple task. Most commonly, investment analysts look to the volatility of returns as the best measure of risk. While this measure has shortcomings, it’s fairly straightforward to determine and has fewer shortcomings than other measures. What analysts mean here is the statistical probability that returns will vary to any given degree from an observed average over a given period of time. Readers familiar with statistics know this measure as “standard deviation.”
To see how we might apply this approach, I looked back over the past six years. Why such a short period? Because that’s when the first index exchange traded fund (ETF) with reliably lower equity volatility first became widely available to investors. This short timeframe has included four distinctly different markets:
• May 2006 though December 2007—rising equity prices with unusually low volatility;
• Early 2008 through March 2009—collapsing equity prices with extraordinarily high volatility;
• March 2009 through April 2010—rapidly rising equity prices with falling volatility; and
• May 2010 through today—very choppy equity prices with medium to high volatility.
The average annualized return for an S&P 500 Index ETF over the past six years comes out to about 2.6 percent. The standard deviation of those returns (measured using daily prices) was just under 24 percent. Applied prospectively, this means that two-thirds of the time, we should expect returns for one year to fall somewhere between +26.6 percent and -21.4 percent, a very wide swing.
In risk budgeting, we’re really only concerned with the downside risk. Assuming downside volatility and upside volatility are the same (and they were close enough for this analysis), for purposes of risk budgeting, we can say that in one year in every 40, we might expect the return to be two times the standard deviation below the average, or a pretty fearsome negative 45.3 percent for our S&P 500 index fund.
Few clients will likely say they can stomach such a drop. But let’s say the client will agree to a one-in-40 chance their portfolio will fall 25 percent in a year. How do we modify the proposed portfolio to achieve that goal? One obvious answer is to apportion the portfolio between stocks and a less risky asset, such as bonds.
Over the same six years, a broad bond index fund had an average annual return of 6.5 percent and a much lower standard deviation of returns of 4.5. Although bonds and stocks are generally not highly correlated, the fact that bonds are much lower in volatility means they reduce our overall portfolio risk much less than most people assume.
First, I evaluated a fairly conservative portfolio composed of 40 percent S&P 500 index fund and 60 percent bond index fund. Given that bonds have outpaced stocks over the past very difficult six years, this actually improved the investment returns modestly from 2.6 percent to 3 percent. However, the volatility of portfolio returns only declined from 24 to 20.9, and our one-in-40 year loss was still almost 39 percent. To bring our risk within our budget of a one-in-40 year loss of 25 percent, I had to adjust the portfolio to be almost 87 percent bonds.
For a better approach to bringing a portfolio within my risk budget, I addressed the volatility of the stock portion directly. I substituted the Vanguard Dividend Achievers ETF (ticker VIG) for the S&P 500 fund. This is an index fund that includes only stocks with growing dividends, much like our firm’s Dividend Growth Portfolio, albeit with less focus on faster growing dividends. A portfolio composed of 38 percent VIG and 62 percent bonds had the same volatility as the one with 13 percent S&P 500 and 87 percent bonds.
As we saw in my May 2012 column [“Volatility to the Fore”], academic research has recently substantiated that, over the long haul, lower volatility stocks provide better returns. Not only were we more effective in reducing our downside risk, our portfolio with 38 percent VIG improved our annualized returns to 5 percent, or 1.2 percent more than the portfolio with equivalent risk using an S&P 500 index fund. Because of compounding, small differences in annualized returns make a big difference in long-term portfolio performance.
