When Backtests Fail

Let’s say you’re building a diversified portfolio and want to include European high yield bonds in the mix. What are you plugging in for an estimated return today?

If you’re like most investors, you’ll look to a backtest to provide the answer, asking the following question: what have the returns been historically?

With data going back to January 1998, European junk bonds have delivered an annualized return of 7.1% (note: total return, using BofA Merrill Lynch Euro High Yield Index). Not too shabby, especially in the era of negative interest rates. We’ll just plug that number in and start collecting on the 7.1%.

Will that work out as planned? Probably not.

Historically, there’s a pretty close relationship between the starting yield and forward returns. In general, the higher the starting yield, the higher the future returns, and vice versa.

In late 2008 when European junk bond yields peaked at over 25%. Investors purchasing those bonds would earn annualized returns over 20% in the next 5 years.

What does this imply for today’s market?

With European junk bond yields hitting an all-time low this week of 2.15% (lower than the yield on 10-year Treasury Bonds), we should expect much lower future returns.

Just how low remains to be seen, as there is no historical precedent for 2% junk bond yields. But barring a change in bond math, it will be nearly impossible for European junk bonds to approach their historical average return of 7.1%. If spreads should widen and defaults pickup, investors would likely be facing negative returns.

When markets hit extremes, backtests are doomed to fail. You can’t assume the future will look anything like the past when the present doesn’t not look anything like the past.

This was true for the Nikkei in 1989, the Nasdaq in 2000, and European junk bonds today. Caveat emptor.