In 1996, Bridgewater Associates launched their All Weather strategy, driven by an approach that would later be called “risk parity”.  Sixteen years later, it seems like just about every firm has launched their own version.

The idea behind risk parity is that traditional capital-based asset allocation actually tends to create a concentrated risk profile.  The frequently quoted statistic is that a 60/40 stock/bond portfolio has 90% of its risk (as measured by variance in returns) explained by equity variance; such a large proportion that our portfolio suffers the potential of permanent impairment which is the real risk that investors care about.

The brilliance of the Bridgewater method was to lever or de-lever asset class returns to a constant risk profile (10%).  This meant that to achieve higher long-term returns, we no longer had to hold a significant amount of equity in our portfolio — we can now use a variety of beta sources.  From there, Bridgewater selects from among the asset classes a portfolio that would perform equally well in one of four economic environments: 1) inflation rising, 2) inflation falling, 3) growth rising, and 4) growth falling.  Economic growth and inflation risk factors are chosen because they are believed to be the primary drivers of asset return variations.

This methodology is rooted in economic reasoning.

Unfortunately, with success comes imitation, and most of the current market offerings are poor imitations at best.  Vanilla implementations use historical estimates of volatility and correlation to construct a portfolio such that each assets marginal contribution to volatility is equivalent.  While this matches the technical definition of “risk parity” (each asset class contributing an equivalent amount of risk), it misses the point.

Consider, for example, the relationship between stocks and bonds.  Stocks are a claim on earnings and are worth more in times of economic growth.  Bonds give a fixed income stream and are worth more in decreasing interest rate or deflationary environments.  Thus, when asset returns are driven by economic uncertainty, we expect them to be negatively correlated.  By levering up bond returns to equal the risk of equity returns, we can put an equivalent amount of capital in each and know that our portfolio is “diversified” to the economic growth factor.  Without leverage, we would likely have had to hold a 20/80 portfolio to achieve equivalent diversification offset, and our long-term expected returns would be much lower.

However, both stocks and bonds are vulnerable to rising inflation!  Therefore, when inflation uncertainty drives asset class returns, we expect them to be positively correlated!  A naive historical estimation may miss this fact.  Bridgewater has stated previously that they do not use correlations in their weighting methodology, but rather weight their assets based on an understanding of how economic conditions are reflected in asset pricing.  Thus, it is important that their portfolio is equally balanced across economic environments.

But people try to cut corners.  They don’t like leverage.  They don’t want to think hard, so they use volatility and correlation estimates to drive their weights.  This means they end up giving those assets with lower volatility higher nominal weights and assets with higher volatilities lower nominal weights and create a very dangerous position: a large concentration in a single asset class.  But didn’t we just say we didn’t care about capital allocation?  Don’t we care about risk budgeting?  Sure: except when your large concentration position continues to exhibit low volatility but has negative returns.  Consider, for example, the 50/50 “risk parity” portfolio in the form of a 20/80 traditional allocation.  In a rising interest rate environment, this portfolio would suffer tremendously.

Our second gripe is that volatility is not what most investors define risk as; rather, they are concerned with the risk of ruin.  While Bridgewater may use volatility as a guide to leverage, they don’t use it as a guide for risk management; rather, they use their economic understanding.  This is where many funds go wrong: they use volatility both as a guidepost for leverage and for risk management.  The risk here is that a straight line down has no volatility.

Finally, the process of using volatility and correlation to drive weights uses asset classes as linear proxies for risk factors.  This is a poor approximation, as asset class  returns are dynamic and non-linear functions of the underlying economic risk factors.  Bridgewater understood this from the very beginning and uses their economic intuition to construct a view of asset pricing relative to economic conditions.  Again, consider the case of stocks and bonds, whose correlation (a statistical summary, not an implicit market characteristic) can swing from negative to positive depending on the driving market factor.  Creating allocations based on volatility and correlation estimates can easily lead to a portfolio that is too heavily concentrated to succeed in only a single economic environment.

Risk parity is a great idea, but we feel it is misunderstood.  The original idea was to construct a portfolio that was equally sensitive to the underlying market risks: economic growth and inflation expectations.  Unfortunately, this has been mistranslated to mean “equal contribution to volatility,” a strategy that heavily overweights fixed income.  Since fixed-income has had tremendous performance over the last decade, people mistake the backtested returns of the strategy for strategy success instead of just a large fixed-income exposure.  However, it makes the portfolio massively susceptible to inflation and rising interest rate environments — the exact opposite of an equal risk exposure methodology!

So caveat emptor!  At Newfound, we think the imitations demonstrate a distinct lack of quantitative integrity; they fail to adequately balance quantitative implementation with economic intuition and open investors to considerable risks under the guise of being “risk allocated.”

Corey is co-founder and Chief Investment Officer of Newfound Research, a quantitative asset manager offering a suite of separately managed accounts and mutual funds. At Newfound, Corey is responsible for portfolio management, investment research, strategy development, and communication of the firm's views to clients. Prior to offering asset management services, Newfound licensed research from the quantitative investment models developed by Corey. At peak, this research helped steer the tactical allocation decisions for upwards of $10bn. Corey holds a Master of Science in Computational Finance from Carnegie Mellon University and a Bachelor of Science in Computer Science, cum laude, from Cornell University. You can connect with Corey on LinkedIn or Twitter.