Monday, March 22, 2021

Changing stock bond correlation and macro factors - what has to happen

 

An expected stock-bond correlation is my primary asset allocation worry and keeps me up at night. So much of portfolio construction has been driven by this negative relationship that a switch to a positive correlation will up-end conventional thinking from the last two decades. 

Simply put, the negative correlation has been the best diversifying hedge in the money management business. If there is a stock decline, hold bonds, get your carry and price appreciation and rest easy. It has been a better alternative than holding the median manager for many hedge fund strategies and it has been cheap diversification. However, the legs may be cut out of this hedge if we have a different inflation regime. There is no carry and if rates rise from inflation, there will be a clear return drag. This would end the benefit of risk parity strategies and end the core benefit of 60/40 base case for portfolio construction. 

A look at longer history shows this correlation has been positive for decades prior to 2000. There was less value with a safety premium during the higher inflation periods. 


The research on this topic has been increasing, so we should be able to anticipate and prepare for the new environment. A good research piece on the macro drivers of this correlation is "Macroeconomic determinants of the correlation between stocks and bonds' by Marcello Pericoli of the Banca D'Italia. This work first shows that there has been a clear and significant change in the stock-bond correlation from positive to negative based on changes in macro relationships. In the US, the growth inflation correlation moved from negative to positive and the real rate inflation correlation moved from positive to negative. This was not just a US macro pattern. A similar pattern occurred in Germany.  


The author breaks down the covariance between stocks and bonds into five components. Of course, the correlation between stocks and bonds will not be the same as the covariance. Higher stock and bond volatility drives the correlation to zero. Some of the components that drive the stock/bond correlation are unambiguous while others are not clear and are based on how inflation is perceived versus growth and real rates. 

First, there is an uncertainty channel associated with real rates which is positive. A change in real rates will have a positive impact on the discounting of both  stock and bond cash flows.  Second, expected inflation will have a negative impact on the co-movement of stocks and bonds. An increase in inflation will negatively affect nominal bond cash flows while stock cash flows will provide a hedge against inflation.  

The next three factor signs are harder to distinguish. They can be called the cash flow channel, the discount factor, and the portfolio rebalancing channel. If the economy is neutral to inflation shocks, then these co-movement is easy to measure. With the cash flow channel, the impact is based on the correlation between stock cash flows and inflation. This correlation is related to  supply and demand shocks and whether inflation is countercyclical or pro-cyclical. The discount factor channel relates to money illusion and whether investors use nominal discounting of real values. The rebalancing correlation is related to the co-movement between dividend yield and real rates and represents the switching around business cycles - negative during expansion and positive during recessions.  This will be associated with the equity risk premium.





What happened to the correlation between stocks and bonds in earlier decades and the current environment? The global economy has moved from a countercyclical expected inflation environment to a pro-cyclical expected inflation environment. This is noted by the big switch in the growth and inflation correlation. 

The stock/bond correlation is driven by uncertainty concerning expected inflation, real rates, the correlation between inflation, dividend and real rates, and equity risk premia. The long-term swings in this correlation are based on the inflation business cycle link, specifically, the link between future cash flows and inflation. A slowdown and higher inflation are not on the top of most minds, but an increase in real rates and growing inflation are occurring right now. 

No comments: