Risk Parity Resources: Philips et al. (2012)

Dynamic Correlations: The Implications for Portfolio Construction

Older paper (2012) but valuable for two reasons: 1) is realistic about correlation, and 2) pays special attention to correlation in times of market decline, such as the one that just preceded the paper and the one we’re going through now.


Read the original:


Important Points for the RP Investor:

This report is put out by Vanguard’s research group for individual investors, and as such, is very accessible, brief, and focused. The authors investigate how correlations between assets have changed over time, with special focus on how they move in times of market turmoil. They track board equity indices versus broad bond indices, though at the end, expand the discussion to include other assets like international equities, REITs, and commodities, among others.

Pages 3 and 4: Goes over the concept of correlation, which you can also read about in my post on the Basics of Risk Parity. Authors note correlation’s importance in a portfolio, but also its limitations. Correlations can change, sometimes subtly but also sometimes abruptly. They note that, over the long-term, US equities and bonds have been only slightly correlated with each other, but with variation depending on the time period.

💡
“While the long-term (since 1926) average correlation between these two asset classes has been 0.25, the figure shows correlations over shorter windows vary widely from this average, with a range of 0.72 for the 5-year period ended 1975 to -0.54 for the 5-year period ended 2005” (page 4).
Source: Vanguard

Pages 6 through 10: Implications for portfolio construction. They argue that evidence shows a positive effect for diversification in a portfolio, but also that, at times of acute market stress, long-run patterns may no longer hold. “During a flight to quality, increased systematic risk tends to swamp asset-specific risk” as all assets can tend towards positive correlation in a downward direction (p. 6).

The authors then focus on two such periods, the 1988-2007 period when a diversified portfolio would have made great sense. Although a 50/50 stock/bond allocation over that time would have returned 9.9% with a standard deviation of 7.4%, the equal weighted six asset-class portfolio (international stocks, emerging market stocks, REITs, commodities, High-yield bonds, and international bonds) would have returned 10.9% with similar vol;atility.Next they look at the 2007-9 crisis, and the same does not hold. The 50/50 doesn’t perform well at all (-26%), but still beats a diversified portfolio by 12%.

On page 10, there is an interesting backtest of various diverse portfolios if only considering their returns in the worst 10% of months for equities since 1988. In these months of diving stock prices, some asset classes tend to follow them downwards, with other ones slightly negative. The one exception is Treasury bonds,  which actually trade above their typical norms when the equity market is struggling.

Source: Vanguard

Pages 10 through 12: Diversification is not only about correlation. Authors write that investors often treat correlation as the only focal point in portfolio construction. They argue that investors should also consider that even if there is positive correlation, assets rarely move in the same direction with the same magnitude, so there are benefits to diversification even when typical correlation patterns change, as they do in market turmoil. They show that even if there is positive correlation, the lower amplitude of a diversifying asset class can still benefit a portfolio and should not be overlooked. Investors would be wise to be patient, as

💡
The benefits of diversification, low correlation, and sensible portfolio construction tend to bear out over longer — 3-, 5-, and 10-year periods, even though they may not be as clear in the very short term (page 11).

In sum, a quick and easy to read paper explaining the benefits of diversification in a portfolio in both good times and bad. Some helpful reminders not to get too attached to a particular correlation number, as they can change when we least want them to. We’re seeing this now, of course, as equities and Treasuries are both in decline, unsettling traditional 60/40 and Risk Parity portfolios alike. RP portfolios tend to have higher allocations to the asset classes that actually have done well lately (namely commodities and managed futures), but I haven’t seen many (any?) with allocations at the level you’d need to fight against this historic simultaneous decline in stocks and bonds.