Credit risk premium ‘underestimated’ due to duration fault, says AQR [updated]
The credit risk-premium generally identified in corporate bonds has been underestimated with research definitively showing its existence and that it acts as diversifier in equity and sovereign bond portfolios.
A time-series analysis research paper from AQR Capital Management showed the excess return of corporate bonds over sovereign bonds had been wrongly accounted for, meaning the risk premium associated with credit risk was not correctly identified.
The paper ’Credit Risk Premium: Its Existence and Implications for Asset Allocation’ also said credit risk premium does exist and would diversify a 60/40 equity and sovereign bond portfolio over the long term.
The research, carried out by Attakrit Asvanunt and Scott Richardson from AQR using data from 1926 to 2014, showed term-risk was wrongly identified as equal in previous studies meaning the actual return allocated to credit risk was underestimated.
AQR said the returns from corporate bonds were driven by changes in the risk-free rate and the spread over the risk-free rate.
To isolate the component of returns attributed to credit risk one must remove the effect of interest rates by subtracting the returns of a risk-free instrument from a corporate bond return, while matching duration.
AQR said previous research isolating credit risk-premium assumed matching duration, but risk-free bonds were generally longer and carried more term premium thus have a higher risk-adjusted retrun than shorter bonds. (see diagram)
Scott Richardson, managing director at AQR, said the average duration was around seven years for a corporate bond but 12 for risk-free sovereign bonds.
This meant estimates of the credit-risk premia were underestimated.
”If you are subtracting the longer-dated government bond, you are subtracting too much of a return from the corporate bond return. And that simple difference pushes down the estimate of the excess credit return,” he said.
“People have effectively over-hedged the corporate bond return; the consequence being that given a positive term premium you are pushing down the estimate of the credit risk premium.”
Source: AQR Capital Management
The paper also found the average monthly credit-risk return was 21 basis points using data from August 1988 to December 2014.
However, the pair also looked at whether exposure to credit risk was beneficial in a portfolio sense, or if the premia was equity risk-premia in disguise.
It tested three portfolios, over different time sets, with different weightings to risk-free government bonds, equities and corporate bonds
A long-only portfolio between 1988 and 2014 weighted 73%, 15% and 12% respectively gave an optimal risk and return profile, while on a longer 1936 to 2014 scale the weightings was 35%, 17% and 48%.
Collectively, the results showed there was a risk-premia to be had from exposure to credit risk which is sufficiently different to equity risk-premia, the paper said.
However, Richardson stressed these were tactical allocation considerations, as different periods of time and economic cycles would yield different results.
He said the finding the existence of credit risk-premia was obvious, but finding the existence of positive risk-adjusted returns for credit was unique.
“You can only uncover the result with careful attention to measurement.”
Earlier this year, AQR teamed up with the London Business School to jointly launched the AQR Institute of Asset Management, forming a 10-year collaboration to fund and generate research across a range of disciplines.
This article, originally published on 13 April, has been updated to correct a misunderstanding by the author