The two most popular junk-bond exchange-traded funds in Canada, iShares U.S. High Yield Bond CAD-Hedged (XHY) and BMO High Yield U.S. Corporate Bond CAD-Hedged (ZHY), represent more than half of total assets invested in the High Yield Fixed Income ETF category. Two large U.S.-traded ETFs that track the same indexes -- 
iShares iBoxx $ High Yield Corporate Bond (HYG) and SPDR Barclays High Yield Bond (JNK) -- garnered a lot of attention earlier this year when activist investor Carl Icahn raised questions on their liquidity and tracking capability. Thus, it's a good time to pull back the curtain and take a closer look.
An important point that has come up time and again in this rivalry is their relative tracking performance. On the surface, there seems to be a large discrepancy between the two ETFs in how well they track their respective benchmark indexes. This discrepancy is visible as much in the Canadian and U.S. versions of these ETFs.
The table below compares the net asset value returns for XHY and ZHY against the returns of their respective benchmarks over different periods. Based on these figures, HYG appears to be a clear standout.
In fact, after adjusting for its 0.67% management-expense ratio, XHY actually outperforms its benchmark on a gross basis. On the other hand, ZHY appears to lag its index by a relatively wide margin--a gap far too large to be explained only by its 0.62% MER.
Even though XHY is pricier than its chief rival, it appears to have tracked its index much better than ZHY. Despite its higher price tag, XHY could actually be considered cheaper overall from a total cost of ownership perspective.
| Tracking difference: NAV performance versus index | |||||
| YTD | 1 Yr | 3 Yr | 5 Yr | ||
 |
|||||
|
|||||
| iShares U.S. High Yield Bond CAD-Hdgd | -0.99 | -2.89 | 3.56 | 5.44 | |
| Markit iBoxx Liquid High Yield TR USD | -0.44 | -2.47 | 3.29 | 5.53 | |
| Difference | -0.55 | -0.42 | 0.27 | -0.09 | |
|
|||||
|
|||||
| BMO High Yield U.S. Corp Bond CAD-Hdgd | -1.59 | -4.04 | 3.01 | 4.91 | |
| Barclays VLI High Yield TR USD | -0.15 | -2.69 | 3.70 | 6.01 | |
| Difference | -1.44 | -1.35 | -0.69 | -1.10 | |
|
|||||
| Source: Morningstar Direct. Returns through Oct. 31, 2015. | |||||
|
|||||
Portfolio check
Given that they track unique benchmarks, there are differences--in terms of portfolio construction and economic exposure--that, of course, should be considered along with cost and tracking efficiency. These factors help explain why performance between XHY and ZHY will vary. The tables below provide a breakdown of some of the key differences between the two portfolios. But in terms of the rules governing index composition, both benchmarks have strenuous liquidity constraints. ZHY's benchmark sets the bar for the minimum outstanding face value of a bond at US$500 million, even higher than the US$400 million limit set by XHY's index.
| Portfolio overview | ||||
| XHY | ZHY | |||
| Assets (M$) | 575.58 | 1,160.18 | ||
| Management-expense ratio (%) | 0.67 | 0.62 | ||
| Distribution frequency | Monthly | Monthly | ||
| 12-month yield (%) | 6.39 | 6.57 | ||
| Yield to maturity (%) | 7.02 | 7.30 | ||
| Duration (years) | 4.09 | 4.16 | ||
| Premium/Discount 12-month avg (bps) | -6 | 3 | ||
| Average daily share volume (3 months) | 37,758 | 57,225 | ||
| Turnover ratio (%) | 29.75 | n/a | ||
|
||||
| Source: Morningstar Direct. Data as of Nov, 18, 2015 | ||||
The funds offer similar exposure to the same asset class. As shown in the tables above, there are some slight differences, but nothing jumps out as significant enough to explain the tracking discrepancy. So what is it? Are iShares' portfolio managers that much more skilled? Perhaps BMO just happens to be lousy at indexing high-yield bonds?
The answer actually lies deep in the fine print. As it turns out, the discrepancy in tracking error may largely boil down to a subtle difference in how the indexes are calculated. 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.
The index ZHY tracks, the Barclays High Yield Very Liquid Index, prices all transactions at the bid price (which has been the industry standard historically). XHY's benchmark, the Markit iBoxx USD Liquid High Yield Index, also marks existing and exiting bonds at the bid price, per the industry standard. But unlike most other fixed-income indexes, it marks new holdings that are entering the index at the ask (offer) price.
This technical detail helps explain why ZHY's tracking error has been so much larger than XHY's. In effect, by pricing incoming bonds on the ask side, transaction costs are embedded in the Markit index that XHY tracks. As a result, XHY should be expected to have minimal (if any) tracking error. In contrast, ZHY would be expected to lag its benchmark by (at least) its transaction costs, given that the Barclays index does not reflect the cost of the bid-ask spread associated with actually buying and selling a particular bond. Theoretically, a portfolio that fully replicates its index, holding for holding, and matches every rebalance and reconstitution should be expected to trail its benchmark by an amount equal to its transaction costs.
Since Barclays also calculates a version of the index that includes transaction costs, it is possible to make a more direct apples-to-apples comparison. The chart below shows the performance of U.S.-traded JNK (which tracks the same Barclays index as ZHY) versus both versions of its benchmark. Compare that against the chart showing the U.S.-version of the iShares ETF against its benchmark index. From this view, both ETFs appear to tightly track their benchmarks.
- Source: Barclays, SSgA, Morningstar Analysts
- Source: Morningstar Analysts
As can be seen in the charts above, once the playing field is leveled, both these funds have tracked their benchmarks reasonably well. ETFs are often judged by how well they track their indexes--but not all indexes are created equal. In this particular case, the devil is in the details.
