Treasury yields and factor returns? Acquaintances, but little in common…
Analysis Date: February 15, 2018
Concerns over rising inflation recently took center stage and for a good reason: strong wage figures, rising producer pricing, and higher-than-expected consumer-price data all pointed to upward pressure on long dormant inflation in the US. The uptick in inflation solidified investors’ expectation that the Fed would raise rates more aggressively in the near future, which, in turn, pushed the yield on the 10-year Treasuries to a four-year high last Friday.
The market has seen few periods similar to the current environment of a rising Fed Funds Rate over the past three decades, raising the question of whether the relationships between style factors and 10-year Treasury yields might shed some light on equity portfolios’ sensitivities to rates.
We analyzed the correlation between returns of style factors in Axioma’s fundamental medium-horizon US4 model and changes in 10-year Treasury yields. Particularly, we calculated the 5-year rolling correlations between weekly factor returns and weekly changes in 10-year yields. Correlations were on average very close to zero over the past 31 years, but the range has been wide. Correlations for the five years of weekly data ended last week are in many cases quite different from zero.
Market Sensitivity reported the strongest sensitivity to 10-year Treasuries, as its returns’ correlation with changes in yields rose to 0.47 on February 15. Market Sensitivity’s returns had been negatively correlated to changes in yields before 2001, but positively correlated afterwards. This is not too surprising, as a regime change occurred in the 2000s, with equities becoming positively, rather than negatively, correlated with interest rates. This change has been attributed to the fact that, in low rate regimes, investors are more concerned about deflation than inflation, so increases in rates are considered a plus.
Volatility experienced a similar pattern as Market Sensitivity, with returns being mostly negatively correlated to changes in yields before 2001 and positively afterwards. Its correlation with changes in yields was 0.32 last Friday.
Correlations of changes in 10-year yields with Exchange Rate Sensitivity have been mainly negative through the 30 years under review, but rose in the past couple of years and reached a three-decade peak of 0.27 this February.
Value’s correlation with yields has been low historically, but as of last Friday it recorded a positive correlation of 0.25, slightly lower than its highest point of 0.38 four years ago.
Although mostly positive in the past decade, Leverage’s correlation with yields has been surprisingly low. As of last Friday, it was slightly below 0.07.
Other factors reporting low sensitivities to yields last week were: Liquidity, Size, and Profitability.
Medium-Term Momentum, Growth and Dividend Yield had inverse relationships with 10-year yields. These factors reported negative correlations in most of the past 30 years. Dividend Yield’s negative correlation was the highest, at -0.24 last Friday.
In summary, correlations today are very different from what they’ve been in the past. Only Market Sensitivity and Volatility’s returns displayed relatively high correlations with changes in 10-year yields in the past five years. None of the other factors has shown a consistently high or low correlation with rates over time, including during similar periods of rising rates.
 10-Year Treasury Constant Maturity Rates
 For a longer discussion on this topic, see “How to Model the Impact of an Interest Rate Rise Using Axioma’s Multi-Asset Class Tool” - especially the section starting on page 14.