The mirage of passive emerging market debt ETFs
Passive ETFs have grown in popularity as a low-cost way to access various asset classes. However, passive emerging market debt (EMD) ETFs have several disadvantages. They are a suboptimal vehicle for accessing less liquid markets, have generally underperformed their stated benchmarks, underperformed versus comparable active managers, and are limited to a constrained opportunity set.
Investors have increasingly allocated money to passively managed ETFs since the vehicle was first introduced to the market in the 1990s. ETFs initially proliferated as a low-cost alternative to replicate equity indices and then in the early 2000s the first passive bond ETFs arrived. In equity markets, the transition from active to passive investment has been well documented; perhaps less commonly observed is that the bond ETF market has also seen rapid growth over the past decade. In emerging market debt (EMD), hard currency passive ETFs catering to the US market have grown in size from $1.6 bn in 2009 to more than $21 bn in 2022.[i]
In principle passive ETFs are designed to offer low-cost exposure to a given asset class. While the advantages of ETFs seem appealing at first glance, they come with some very important tradeoffs. When taking a closer look at passive EMD ETFs, one finds that they often miss the mark of their stated goals and entail disadvantages when compared with an open-ended, actively managed fund. Specifically, we argue that passive EMD ETFs:
- Are a suboptimal vehicle for accessing less liquid markets like emerging market debt
- Have generally underperformed their stated benchmarks
- Have underperformed versus comparable active managers in an asset class where inefficiencies still exist
- Are limited to a constrained opportunity set and forgo exposure to a significant portion of the opportunity set.
ETFs are not the optimal structure for accessing less liquid markets
Unlike open-ended mutual funds, an ETF trades throughout the day at a discount or premium relative to the underlying value of its holdings (NAV). Oftentimes the difference between the NAV and market price is minimal because shares are liquidated or created by authorized participants. Authorized participants effectively buy the ETF’s underlying securities and bundle them into ETF shares. Therefore if a material difference between the NAV and ETF price exists, authorized participants are incentivized to arbitrage that difference away at a profit.
This process works best with stocks, which are highly liquid, but the process is not as ideal for fixed income ETFs. This is especially true for segments of the fixed income market that are less liquid, such as emerging market debt. In the equity market there is often just one instrument for each issuer, which has the dual effect of bolstering liquidity and minimizing deviations in ETF premiums and discounts. However, in bond markets corporates and sovereigns often issue a range of bonds with different maturities and coupons. For instance, the EMBI Global Diversified Index has approximately 160 issuers but over 900 securities.[ii] Because bond markets contain a much larger number of securities, during periods of market stress, liquidity in lesser liquid markets is much more likely to dry up, resulting in little-to-no trading in some instruments. Therefore ETF investors may be handicapped compared to an open-ended fund holder, who can buy and sell daily at the NAV. According to Bloomberg, during the peak Covid stress roughly 70 fixed income ETFs traded with at least a 5% discount to their NAV and 16 traded at a discount of 10% or greater.[iii] This represents a material cost for investors looking for liquidity when it is needed most.
Significant underperformance versus the named benchmarks
A key goal of a passive ETF is to provide a reasonable performance proxy to the underlying asset class. When analyzing the passive US EMD Hard Currency ETF cohort, we see that on average these funds have failed to meet their objective. The chart below tracks the underperformance of the passive ETF cohort relative to their stated benchmarks. As you can see, this underperformance has been relatively consistent and significant.
Figure 1: US EMD ETFs generally underperform their respective benchmarks
Source: Morningstar as of 6.30.2023. Average returns based on a simple average of USD denominated Passive ETFs and benchmark returns for the periods listed. Past performance provides no guarantee of future results. Indices are unmanaged. Actual results would vary due to, among other things, fees and expenses.
Underperformance versus comparable active managers
A highly touted advantage of passive ETFs is their lower fees when compared to active managers. According to our analysis of the passive EMD ETF peer group, the average expense ratio charged by passive US hard currency EMD ETFs is 37 bps compared with 102 bps for US EMD open-ended funds. While equity ETFs are often priced much lower, EMD ETFs’ small cost advantage often does not offset the value added through active management.
As shown below, despite modestly higher expense ratios, the average active EMD fund manager has outperformed the comparable ETF by a comfortable margin net of fees.
Figure 2: Active managers outperform across most time periods
Source: Morningstar as of 6.30.2023. Average returns based on simple average of USD denominated passive EMD ETF and EMD active manager returns for the periods listed. Active managers defined as the US Emerging Market Bond Category (Ex-ETFs) for USD denominated share classes. Past performance provides no guarantee of future results and may not be repeated.
Poor asset class coverage
Given the large size of most passive ETFs and their mandate of tracking their index, they are generally managed against indices that do not fully capture the market segment as advertised. The three largest passive hard currency EMD ETFs: iShares Emerging Markets Bond (“EMB”), Vanguard Emerging Markets Government Bond (“VWOB”) and Invesco Emerging Markets Sovereign Debt (“PCY”), have $15.6 bn, $3.5 bn and $1.6 bn in assets, respectively (as of June 30, 2023). These three strategies command more than 90% of the hard currency EMD ETF asset base. All three strategies are benchmarked against dollar-denominated sovereign bond indices, forgoing two thirds of the asset class, corporates and local currencies.
The largest ETF, JP Morgan’s EMB ETF, is benchmarked against the JP Morgan EMBI Global Core Index, which has a market cap of $580 billion. This opportunity set is more limited than the more commonly used sovereign index, the JP Morgan EMBI-Global Diversified Index, and equates to just 11% of index-eligible emerging market debt. Likewise Invesco’s PCY ETF utilizes a constrained index that is limited to securities with at least three years to maturity and at least $500 million outstanding. Furthermore, this index eschews quasi-sovereign bonds entirely. Quasi-sovereigns, entities that are owned or guaranteed by sovereigns, are often an attractive and higher yielding alternative to sovereign bonds with fairly similar risk profiles. Quasi-sovereigns make up nearly 20% of the most common sovereign market index, the JP Morgan EMBI-Global Diversified Index. These overly constrained indices used by many passive ETFs can often conflict with an investor’s objective of obtaining a representative return of the asset class.
In contrast to the above limitations, most active managers have a much broader opportunity set in which to invest. This often means that active managers can allocate to more diverse issuers that frequently offer an adept active manager the opportunity to take advantage of market inefficiencies. Additionally, active managers have the ability to add value by investing in off-benchmark securities and new issues that are often absent from the ETF investment universe.
The chart below demonstrates that this is an asset class that has provided opportunities to generate excess returns. Over the past ten calendar years, the difference between the highest and lowest performing sub-segment of EMD was greater than 4% in 8 of those 10 calendar years.
Figure 3: Annual Returns of EMD Sub-components
Source: JPMorgan and Bloomberg. Sovereigns are represented by the JPMorgan EMBI Global Diversified Index; Corporates are represented by the JPMorgan CEMBI Broad Diversified Index and Local Currency is represented by the JPMorgan GBI-EM Global Diversified Index. Past performance is no guarantee of future results. Actual results will vary.
Ultimately, given their poor asset coverage, suboptimal vehicle structure and failure to meet their performance objective, we believe that ETFs are a particularly poor choice for investment in the EMD asset class.
i Morningstar Direct
ii JP Morgan EMBI Monitor
iii Bloomberg.com “Bond ETFs Survived 2020 Liquidity Scare. But Just Barely.”