“The ‘new normal’ in asset allocation must be forward-looking and driven by macroeconomics, said Sébastien Page, Global Head of Client Analytics, Executive Vice-President at PIMCO. He was addressing a CFA Society Toronto luncheon on October 15, 2012 in Toronto’s historic National Club.
Traditionally, asset allocation focussed on diversifying according to asset class. In the ‘new normal,’ Page recommends diversifying across risk factors. “Think of asset class as simply a container of risk factors,” he suggested. He gave another metaphor in line with the luncheon crowd. “Think of risk factors as components like proteins, carbohydrates, and fats. An asset class would then be a meal, with different meals having different ratios of components.”
Another aspect of the ‘new normal,’ said Page, is that a static approach to asset allocation, where it is revised every 3 to 5 years, is no longer sufficient. He called for an asset allocation that focusses on the secular and the cyclical investment horizon. [Ed. Note: Here, “secular” means “occurring or persisting over an indefinitely long time.”]
In traditional asset allocation, volatility is used as the sole risk measure. However, the ‘new normal’ must explicitly seek to hedge fat tail risk, said Page. Volatility and fat tail risk are different entities, he reminded the audience, citing an example of two curves with equal Sharpe ratios but one having double the tail risk. “Volatility does not capture the asymmetry.”
Page touched briefly on best practices in alternative asset classes such as real estate, hedge funds, private equity, and specific investments.
The most startling slide Page showed was a data work-up which had miniscule cross correlations between risk factors (3 to 4 percent) whereas the cross correlations between the asset classes were a whopping 30 to 60 percent. The results did not depend on whether the cross correlations were calculated during quiet, turbulent, or in-between periods. It was the same trend throughout: hands down, risk factor cross correlations were much smaller.
“Asset class diversification is not the same as risk diversification,” Page said. He showed companion pie-charts for a portfolio. One pie-chart was segmented according to asset class (global equity, emerging markets equity, energy, venture capital, etc.). Its companion pie-chart showed segmentation according to risk factors (developed markets equity risk, muni spread risk, high yield spread risk, emerging markets currency risk, emerging markets equity risk, etc.)
Page presented a simple empirical model for scenario analysis. The factor return was a linear combination of “GDP surprise” and “inflation surprise.” In this context, “surprise” is the difference between the realized value and the market forecast value of the quantities (gross domestic product and inflation).
When it comes to beta, what matters, according to Page, “is the surprise.” Expectations tend to lag realizations due to the “looking over the shoulder” bias.
Page showed sensitivity analyses for his simple empirical model. These were 2-dimensional colour contour maps, with GDP surprise on one axis and inflation surprise on the other. Bonds showed features on the contour map that were in general opposite to the effect of equities.
Finally, Page covered explicit hedging of tail risk. “Tail risk hedging is the ultimate diversification,” he said. There might be a feedback component to this, he noted. As an analogy, Page noted that race-car drivers have the safest cars. It is the very safeness that allows them to take proper risks. Similarly, de-risking a portfolio could lead to gains in profit-making activities. ª
The research paper “Asset Allocation: Does Macro Matter?” can be found at: https://canada.pimco.com/EN/Insights/Pages/Asset-Allocation-Does-Macro-Matter-Part-II-.aspx