Tag Archives: volatility-adjusted returns

Are Corporate Bonds Worth The Risk?

Reprint of my latest article for Citywire —-

What are the benefits of adding exposure to investment grade corporate bonds (IGCs) in a stock and bond portfolio?

It may sound like a simple – possibly even stupid – question, but new research suggests that IGCs might not be all they’re cracked up to be. You might even be better off sticking with plain old Treasurys.

The conventional wisdom is that IGCs should improve a portfolio’s volatility-adjusted returns. After all, corporate bonds typically have higher coupons than government bonds without having completely similar trading patterns.

Unfortunately, according to recent research by Jared Kizer of Buckingham Asset Management, the yield premium of IGCs over government bonds doesn’t add much at all to the risk-adjusted returns of a stock-and-government bond portfolio.

In a new paper, Kizer has found that the supposedly attractive premium is based on the historical numbers of the early part of one data set when, especially during the Great Depression, significant credit stress delivered high returns to investors for very highly rated corporate bonds.

Kizer went on to question the veracity of that early data, showing that in later periods – the past 50 years – corporates haven’t added risk-adjusted benefits to performance. We will come back to that dodgy dataset shortly, but for now let’s focus on the past half century.

Examining the period from 1969 to October 2017 using Ibbotson data spliced with some from Bloomberg, Kizer found a compounded return premium of IGCs over government bonds of 1.1% per year. By itself, that is statistically significant. However, Kizer also found that the return premium can be explained by different equity factors, including size, value, momentum and others.

Then, using just the Bloomberg data from 1973 to October 2017, Kizer identified an IGC premium that isn’t statistically significant even before accounting for the equity factors. Basically, adding corporate bonds to a portfolio of stocks and government bonds adds nothing to the volatility-adjusted returns of those portfolios.

Next, as he did in a different piece of research into the diversification benefits of real estate investment trusts with Sean Grover, Kizer tried to replicate the performance of IGCs in portfolios constructed without them. Using four simple portfolios of capitalization-weighted stock indices and government bonds, Kizer found that he was able to replicate the performance characteristics of corporate bonds, concluding that ‘corporate bonds are redundant in portfolios that own stocks and government bonds.’

Misdirected affection

But if this is so, why have others thought that corporate bonds added diversification?

A previous research paper by Attakrit Asvanunt and Scott Richardson touted the benefits of IGCs, but as Kizer explained, the evidence that Asvanunt and Richardson marshalled was unduly influenced by the period around the Great Depression, when corporate bonds might have added diversification. Kizer also doubted the accuracy of some of the early period data.

Specifically, Kizer found a significantly higher Sharpe ratio for the IGC premium compared with either equities or interest rate risk during the period encompassing the Great Depression. Anomalous high returns during the 1930s, a period of significant credit stress and an IGC premium in excess of 2.5% per year in this early period for an index focused on AA and AAA corporate bonds resulted in an IGC premium that seems to be completely disconnected from the frequency of corporate bond defaults.

Directing his analysis to the period between 1930 and 1968 in the Ibbotson data, Kizer found that the IGC premium for that period was ‘more than two times higher than market premium and almost five times higher than the term premium over the pre-1969 period.’ The IGC premium itself was 2.6 % per year, which is high considering that the Ibbotson data is oriented toward the highest quality (AA and AAA) corporate bonds. What’s more, Kizer found that the Sharpe ratio of the IGC premium was well in excess of one in the 1930s and 1940s, but never one or higher in any other decade through 2009. The IGC premium had the highest risk-adjusted returns in three of the four
early-period decades.

Dodgy data

All of this is suspicious because default rates were highest in the 1930s, according to Moody’s data. Somehow, the IGC premium was highest when defaults were highest. Moody’s also reports almost no corporate defaults from the early 1940s through the 1960s, meaning that one might expect the 1940s to have produced a higher IGC premium than the 1930s. Although Kizer wasn’t sure, he suspected that the early data tracked bonds that were rated AAA or AA, but then dropped them if they were downgraded or if they defaulted. That means that their poor performance might not have been accounted for accurately.

The upshot for investors is that adding a permanent allocation of investment grade corporate bonds to a portfolio of stocks and government bonds does not increase that portfolio’s risk-adjusted returns. Kizer did allow for the fact that it might be possible to add corporate bonds opportunistically, or when spreads are relatively wide, reducing exposure to them again when spreads contract. This could theoretically enhance a portfolio’s risk-adjusted returns.

Of course, just as we might be living in a period of higher equity valuations, we might also be living in a period of tighter spreads. That would make such an operation difficult to execute and perhaps not worth the risk premium – or the lack thereof.