“The discrepancy in tracking error may largely boil down to a subtle difference in how the indexes are calculated,” Gabriel said. “To be clear, I’m not referring to the composition of the benchmark (that is, which bonds are included and at what weighting), but rather how bonds in each index are priced. The price at which a bond enters or exits the index is a critical detail. The same goes for when existing constituents are bought and sold in the monthly rebalancing process.

David Mazza, head of research for ETFs at State Street Global Advisors, attributes JNK’s underperformance to transaction costs. When an ETF rebalances its portfolios to better track the underlying index, the fund must buy or sell debt securities. While iBoxx absorbs some of the transaction costs, Barclays indices does not, which may help explain JNK’s poorer performance to its benchmark. [The Hidden Costs Of Junk Bond ETFs]

JNK’s underlying Barclays High Yield Very Liquid Index prices all transactions at the bid price, whereas HYG’s Markit iBoxx USD Liquid High Yield Index marks existing and exiting bonds at the bid price and new holdings that enter the index at the ask or offer price.

“In effect, by pricing incoming bonds on the ask side, transaction costs are embedded in the Markit index that HYG tracks,” Gabriel added. “As a result, HYG should be expected to have minimal (if any) tracking error. In contrast, JNK would be expected to lag its benchmark by (at least) its transaction costs, given that Barclays index does not reflect the cost of the bid-ask spread associated with actually buying and selling a particular bond.”

Once investors compare apples-to-apples by calculating both indices by similar transaction costs, both funds have actually tracked their benchmarks reasonably well, according to Morningstar data.

For more information on the speculative-grade debt market, visit our junk bonds category.

Max Chen contributed to this article.