A comparative analysis of CLO ETF returns

Exchange-traded funds (ETFs) composed of collateralized loan obligations (CLOs) have grown significantly in recent years. The underlying CLOs consist primarily of senior secured broadly syndicated loans (BSLs). Although retail investors have had access to BSL funds for over 30 years through open-ended leverage loan mutual funds and for over a dozen years through ETFs, the first publicly traded CLO ETF was launched only in 2020. Prior to this development, CLO investments were predominantly made by institutional investors. This is particularly noteworthy given that historically anywhere from a half to up to two-thirds or more of the US BSL market are held by CLOs[1].

This paper examines the performance of a sample of publicly traded CLO ETFs based on historical returns. The sample encompasses various CLO ETFs that target a range of tranche seniorities, reflecting different levels of credit risk as indicated by their ratings. Additionally, the performance of BSL ETFs, from which we selected a sample, is considered to provide a reference point, given that a CLO tranche is fundamentally a derivative of its underlying BSL portfolio.

Our analysis of the historical daily returns and correlations of the ETFs indicates that over a longer period (such as two years), the risk/return characteristics amongst the ETFs are generally consistent with the underlying risk profile of the relevant ETF – i.e. similar performance levels for similar risks – and moderate correlations. However, an examination of daily CLO ETF returns and correlations over a short volatile period can exhibit a noticeable divergence in absolute and relative performance. These findings indicate that the ‘intuitive’ view that CLOs and BSLs are highly correlated may not be evident until there is significant market volatility. And even in that case, differences in performance indicate that other factors may be at play.

Our analysis found that CLO correlations may be further explained at times by vintage and underlying asset manager exposure rather than just broader BSL market dynamics. We offer additional insights and key factors that can impact the performance between CLO portfolios.

Historical Returns – Data

Our sample of CLO ETFs spans the range of tranche seniorities and the credit ratings scale – from CLO ETFs that focus primarily on senior tranches (rated primarily Aaa) to those that focus on investment grade mezzanine tranches (rated from Aa to Baa). And even to those that include some speculative grade tranches (Ba). Our study also considers BSL ETFs to acknowledge that CLOs are derivatives of the BSL market.

In this context, the relationship between CLOs and the underlying BSL market could provide additional insight into the performance of CLOs. We also included some other market related ETFs to gain additional insight as to the performance of the BSL and CLO markets relative to the broader capital markets. Thus, the selected ETFs can be grouped into three categories: BSL, CLO and other broader markets.

BSL Category

We selected four BSL ETFs that are managed by well-established asset managers and have benchmarks to broad BSL indices.

Table 1: List of BSL ETFs

CLO Category (BSL-Derived)

We selected a variety of CLO ETFs that invest in CLO tranches across the CLO capital structure, and with investments in CLOs across a range of vintage periods and asset managers.

Table 2: List of CLO ETFs

Understanding the underlying ratings distributions gives further insight into the level of exposures a CLO ETF has within a typical CLO capital structure. The CLO portfolios within our sample cover a range of credit risk exposures – from a fund with essentially 100% Aaa to other funds with a broader representation of investment-grade ratings as well as funds with concentrations of Baa/Ba credit risk[2].

Other Broader Markets Category

We also included ETFs that can provide additional perspective on the relative performance of the BSL and CLO markets to the broader capital markets.

In this context, we selected a small cap equity ETF (IWM) since many BSL borrowers would fall in the small cap category; a high yield bond ETF (HYG) since these issuers have a similar credit risk profile (though with different recovery rate and interest rate risks) as compared to BSLs; and a short-term Treasury fund ETF (VGSH) to provide some benchmark of shorter-term risk-free interest rates.

Looking at Financial Performance

We focused on the risk-return and correlation characteristics across the ETFs. We looked for insights into how the ETFs performance behaved with one another within the same category (‘intra-category’), and across categories (‘inter-category’) over different time periods. We observed that both intra-category and inter-category performance and correlation metrics depended heavily on which period we selected. We found that during the most recent market volatility (April 2025), patterns emerged that were not so evident during calmer periods.

In some cases, the correlations we observed between the BSL and CLO ETFs were not as consistent as expected. Understanding that CLOs are derivatives of the BSL market, we initially expected to see a relatively higher degree of correlation across all market environments, but the correlations across all market environments varied. This observation suggests that not all CLOs track the same broad ‘BSL market’.

We also found that CLO performance could be influenced by unique factors related to range of vintage periods of when the underlying CLOs were issued and the overall exposures to common CLO asset managers.

These findings show that these variations across CLO portfolios can offer an opportunity for CLO investors to diversify their BSL and BSL-derivative portfolios to better optimize their specific risk-return objectives. In other words, not all BSL and CLO ETFs are alike.

Historical Performance – ‘Normal Times’

We chose two distinct historical return periods to begin our comparative analysis. One that we could classify as a ‘normal’ period and the other as a ‘volatile’ one. We chose the month of April 2025 as the volatile period. This period reflected significant market volatility due to the rapidly changing global trade outlook and economic uncertainties associated with the related US tariff announcements.

We selected the two-year period from April 2023 through March 2025 as a proxy of ‘normal market conditions’ and can be considered somewhat as a benchmark. Exhibits 1 and 2 show the historical return statistics and correlations of daily returns for each of the ETFs, respectively. Note that two of the ETFs in our sample were not in existence for the full two year-period analyzed. Thus, summary stats are not available, and correlation stats apply only for the respective period that each of these two ETFs was in existence.  

We note that the return and risk characteristics across the ETFs are generally as expected within each investment category and subcategory. For example, among the BSL ETFs we note that there is no material distinction among the standard deviations and coefficients of variation. Among the CLO ETFs, the riskier CLO ETFs with lower rated tranches show higher volatility and coefficients of variation as compared to those CLO ETFs with higher rated tranches. The high yield bond ETF was the most volatile among the credit-sensitive ETFs. As expected, the small cap stock ETF was the most volatile while the short-term Treasury ETF was the least.

During ‘normal times’ NONE of the observed correlations were assessed to be Strong or Very Strong. All the correlations were assessed to be in the Moderate and Weak/Very Weak categories. Only 4 of the 22 observed correlations were assessed as Moderate while the remaining 18 were Weak / Very Weak. Three of the four that were Moderate, were related to the BSL category – BSLs to: BSLs, HY Bonds, and small cap stocks, respectively.

These observations suggest that during normal times, the correlations were not particularly significant for the CLO ETFs. Furthermore, CLO performance did not appear to be materially correlated to other credit risk assets, including the BSL market. Even within the CLO category, correlations were relatively marginal across the CLO capital structure, which suggests that movements in CLO tranche risk premiums seem to be more idiosyncratic during stable markets. These results are not surprising given that the CLO tranches are supported with credit enhancement – unlike CLO equity tranches, which we would expect to be more sensitive.

Exhibit 1: ETFs historical return statistics (April 2023 – March 2025)         

Exhibit 2: ETFs Historical Daily Return Correlation Matrix (April 2023 – March 2025) 

Exhibit 3: ETF Categories Historical Daily Return Correlation Matrix (April 2023 – March 2025)

Historical Performance – ‘Volatile Markets’

As previously explained, we defined the ‘volatile’ period to be the month of April 2025, a period of significant market volatility. Exhibits 4 through 6 show the various historical return statistics for the ETFs during this period.

One can immediately notice the significant jolt in the related risk statistics for all ETFs and the correlations among them. Below are some noteworthy observations based on the comparison of risk statistics between the two periods.

• CLO ETFs experienced the largest increase in risk measures, measured both by volatility (4x to 11x increase) and the range of daily returns (1.5x to 4x higher).
• BSL ETFs experienced roughly a 4x increase in volatility and about a doubling of the range of daily returns.
• Traditional ‘risk’ asset categories of high yield corporate bonds and small cap stocks did not show as much of a relative increase as the BSLs and CLOs – a relatively modest 2x increase in volatility and a 50% increase in range of daily returns. These asset classes, albeit riskier as they are typically subordinated relative to BSL, are more liquid and established markets. For the short-term Treasury ETF, the risk performance in April 2025 was relatively indistinguishable than during the ‘normal’ period.

With respect to correlations, our observations show a sharp increase. Whereas during ‘normal’ times, correlations are not meaningfully significant, this changed during ‘stressed’ markets –  half of the observed correlations are now assessed to be Strong and Very Strong (11 of the 22 observations) whereas 3 of the 22 are now Weak/Very Weak.

While correlations increased substantially, there were still some areas of divergence in performance that are worth noting. For example, the CLO ETFs such as the higher rated investment grade CLO ETFs show moderate correlations to all other asset classes. This implies that CLO ETFs may offer some diversification benefits (especially as you move up the capital structure associated with greater credit enhancement) even in stressed markets. However, the CLO ETFs with lower-rated tranches did show very strong correlations to the ’risk’ assets, but that is to be expected given their tranches’ higher degree of credit risk exposure associated with lower levels of credit enhancement.

Exhibit 4: ETFs historical return statistics (April 2025) 

Exhibit 5: Historical Daily Return Correlation Matrix (April 2025)

Exhibit 6: ETF Categories Historical Daily Return Correlation Matrix (April 2025)

Other Observations

A detailed attribution analysis of the underlying CLO ETF performances is outside the scope of this paper. However, we performed some high-level reviews of the CLO portfolios to look for potential factors that could affect the performance of the CLO ETFs and help explain some of the correlation behavior CLOs experienced, especially during the ‘volatile’ period as CLOs appeared to exhibit lower intra and inter correlations than the other asset classes.

Additional factors beyond market considerations appear to emerge when we look closer at the CLO portfolios. In addition to diverse capital structure exposures, we observed that the CLO ETFs had relatively diverse CLO vintage exposures. While exposures to common asset managers across the CLO ETFs could be significant, we noticed that the exposures were often across various CLO vintages of common managers.

Investing across CLOs with unique asset managers can provide diversification benefits despite targeting the BSL market. Different managers may have varying investment styles, strategies, size (or AUM), as well as industry/sector and credit expertise.

Notwithstanding the fact that a CLO portfolio may consist of several CLOs managed by the same entity, there can be advantages from diversifying across vintages. CLOs can be executed under different market conditions and potentially with variations in reinvestment criteria, even given identical CLO portfolio managers. Furthermore, CLOs from different vintages may be at various stages in their lifecycle, such as the reinvestment or amortization period, which also affects the level of a CLO’s reinvestment activity, as well as being past their non-call period, which can indicate the degree of potential refinancing activities.

Conclusion

ETFs composed of CLOs, with underlying BSLs, have experienced significant growth in recent years. While retail investors had access to funds of BSLs for over 30 years, the first publicly traded CLO ETF was introduced in 2020. This is noteworthy since a large part of the BSL market is held by CLOs and thus offers a broader group of investors to participate in the BSL-derived market across various risk/return profiles.

This paper analyzed the performance of CLO ETFs based on a sample of historical returns. The sample included different ETFs that target a range of CLO tranche seniorities, representing varying levels of credit risk. The performance of some BSL ETFs was also reviewed since the performance of a CLO tranche is essentially derived from its underlying BSL portfolio.

Our analysis of the historical daily returns and correlations of the ETFs show that over a longer ‘normal’ period, co-movements in performance are moderate and the risk/return characteristics across the ETFs are generally consistent with their underlying risk profile. However, CLO ETF returns and correlations can exhibit a noticeable divergence in performance and increase in volatility over a shorter period of extreme uncertainty. An example of which was during the recent period of market fluctuations caused by the US tariff announcements.

We also found that CLO correlations can be explained further by key factors such as the distribution of exposures to: (1) seniorities of the CLO tranches, (2) the vintage periods of when the CLOs were issued, and (3) asset manager overlap within a CLO ETF portfolio.

The bottom line is that CLO ETFs appear to offer investment diversification benefits and that not all CLO ETFs are the same, even given similar credit risk. While performance may appear to converge during stressful times, key differences in performance is also evident.

---

[1] Guggenheim Investments research dated December 7, 2023 – ‘Understanding Collateralized Loan Obligations’ (https://www.guggenheiminvestments.com/perspectives/portfolio-strategy/understanding-collateralized-loan-obligations-clo) and FT Opinion On Wall Street dated June 7, 2025 – ‘The trend strengthening the hand of big credit houses’ (https://www.ft.com/content/b5693e95-3a89-4df0-a1f8-ad3e70338458).

[2] Although some of the CLO ETFs may report CLO tranches that are “not-rated”, this is not technically correct in all cases since some ETFs generally report ratings only from the two largest and widely recognized NRSROs. Nonetheless, “non-rated” tranches may still be less liquid and more volatile given the absence of a rating from one of the two largest NRSROs, but likely to offer higher yields.