How to Minimize Portfolio Volatility and Sleep (a lot) Better

That an investor is able to sleep well at night matters. It matters first for peace of mind, since the sickening feeling of a threatened retirement resulting from plunging markets is not what life is all about. It matters also to prevent the rash reaction of selling out at precisely the wrong time, after a plunge and before recovery has come about.

Many portfolio structures that are described as conservative or balanced are actually quite volatile. Let's explore how to adjust a portfolio to reduce volatility. We'll use two free online tools - the Stingy Investor Asset Mixer, which provides long term return data of the main asset classes, and the recently launched InvestSpy Calculator, which provides market price volatility data. Both allow the user to enter various combinations and percentage allocations over different time periods, a very useful feature for "what if" testing.

Test Portfolio - The Simple Recipe Portfolio
We'll examine a classic simple portfolio, one of a bunch of such portfolios we compared here. The Simple Recipe has only four ETF holdings, three equity (Canadian TSX Composite ETF: trading symbol XIC, USA S&P 500: symbol SPY and international developed country MSCI EAFE: symbol EFA) and the Canadian bond universe (both government and corporate): symbol XBB. Assuming a 50 year old investor, the Simple Recipe allocates 50% to XBB, 25% to XIC, 8% to SPY and 17% to EFA. Those are the numbers we entered into the two tools to get the results of the Pre-Adjustment version of the portfolio.

Reduce volatility by equalizing the Risk Contribution of each holding
Beginning with the InvestSpy tool we enter the initial Pre-Adj allocations and find that almost all the volatility in the portfolio, as seen in the Risk Contribution result column, comes from XIC and EFA. XBB provides a powerful offsetting negative volatility reduction (see screen shot of results below).

(click image to enlarge) 

The source of this effect is revealed in XBB's negative numbers in the correlation matrix.

Next we reduce the overall portfolio volatility by changing the allocation to each ETF to try equalizing the Risk Contribution. That means boosting XBB's percentage and reducing both XIC and EFA. Remember that playing with the numbers is not a matter of trying to find the exact perfect minimum volatility. There really isn't such a thing as an optimal solution, since what works best for the past one year goes out of kilter for the past two years, five years or the entire price history. Volatility and correlation has changed over time and will continue to do so in future. In addition, SPY and EFA are traded in US dollars so that currency shifts with the Canadian dollar will alter the results. We are merely looking for something more stable than the initial portfolio, knowing that it also won't be perfectly adapted to the future. The before and after-adjustment results are summarized in the table below.

(click image to enlarge)

Our Volatility-adjusted allocation reduced volatility by about half compared to the initial portfolio allocation! That includes the period of the financial crisis in 2008-2009. One big takeaway is therefore - increasing the bond allocation brings much stability to the portfolio.

What happens to the returns, do we lose half the returns too?
To find out we entered both pre- and volatility-adjusted allocations in Stingy Investor for various time periods. The Stingy tool has the merit of taking account of currency shifts by converting returns to Canadian dollars, i.e. we assume that the investor does not hold a CAD-hedged ETF. The negative numbers in the Alpha input column approximate ETF expense ratios (we entered current MERs for our ETFs), which reduce returns. This is necessary to get a reasonable estimate of what would have happened since Stingy does not use actual ETF data. ETFs did not even exist in 1970. Looking at the longer historic time data helps us see how our portfolio allocation would have fared through more economic and market environments like the high inflation 1970s oil crisis and the 2000 Tech bubble. Unfortunately, all the data is only year-end annual so we cannot see what happened day-to-day during any year. However, it's better than nothing. We online investors must make do with what we can get, knowing in any case that we are always approximating, since the future is never exactly like the past.

The results vary slightly amongst sub-periods and the total time period for which data is available, 1970 to 2012, but a pattern is clear.

The screenshot below shows detailed results for the volatility-adjusted portfolio:

(click to enlarge)

We've taken the results and created the following summary table comparing the pre- and volatility-adjusted portfolio allocation results for the total time period 1970-2012 and the high inflation years of the 1970s, when bonds might have been thought to severely drag down the portfolio performance:

(click to enlarge) 

• Returns for the volatility-adjusted portfolio are still highly positive but reduced by about a half-percent a year BUT
• Downside risk is hugely reduced - fewer downside years, a lot less volatility and especially, worst drop years have very small decreases

Bottom line: It is quite possible to drastically reduce sleep-depriving portfolio volatility without much loss in returns. Is that a worthwhile trade-off? It's up to you the investor to decide.

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.