Executive Summary:
- Asset-based finance (ABF) focuses on tangible assets such as residential mortgages, consumer finance, and data infrastructure. Our research, using a public market equivalent (PME) approach finds that the risk-adjusted returns of ABF are attractive and exhibit lower correlation to equities compared with corporate direct lending, covering the market performance from 2000 to 2024.
- Estimated returns from ABF are higher than those from direct lending, as we believe this asset class offers better valuations and a larger liquidity risk premium.
- We propose a risk factor framework to effectively quantify the benefits and risks of ABF, understand its interplay with existing portfolio assets, and determine optimal allocation sizes.
- ABF may support portfolio stability during economic fluctuations due to its less-cyclical return profile, which can capture a diverse range of risk factors with lower correlations to traditional asset classes, making it a good complement to corporate debt exposures.
Asset-based finance (ABF), which refers to investments secured by tangible assets, is gaining traction among institutional and wealth management investors. As traditional bank lenders retreat amid tighter regulations and financing conditions, nonbank lenders are capitalizing on opportunities in sectors such as residential mortgages, consumer finance, aviation, and data infrastructure. With a market value estimated in excess of $20 trillion, ABF offers a compelling opportunity for private credit investors. (See our December Alternatives Investment Outlook, “Private Credit: Asset-Based Finance Shines as Lending Landscape Evolves.”)
In our view, the advantages of ABF are that it has the potential to enhance diversification, provide resilience in fluctuating economic conditions, and deliver attractive returns. However, many investors lack the robust framework necessary to effectively measure these advantages. Without such tools, assessing ABF risks, understanding its relationship with existing portfolio assets, and determining optimal allocation sizes become significant challenges.
This paper introduces PIMCO’s public market equivalent (PME) benchmark and risk factor model for ABF funds, which provide the tools to evaluate performance, measure risks, and determine optimal allocation strategies for effective portfolio construction.
Looking back: The historical risk-return profile of ABF is attractive
Our PME benchmark adjusts public securitized fixed income returns based on the credit quality, leverage, interest rate exposure, and spread duration typically associated with ABF funds. Given the higher average credit quality and lower unleveraged yield of ABF investments compared to direct lending funds, we assume greater leverage in private ABF funds. We use 1.25x leveraged public asset-backed securities (ABS) and residential mortgage-backed securities (RMBS), alongside 0.75x leveraged bank loans as PME proxies for ABF and direct lending, respectively.1
Our historical analysis of PME returns focuses on beta returns, excluding additional returns that private strategies might generate over public markets.2 Exhibit 1 illustrates the risk-adjusted returns of the ABF beta, which were particularly attractive in the post-GFC era on a standalone basis and relative to direct lending beta.
Exhibit 1: Historical analysis of risk-adjusted returns for ABF and direct lending PMEs
SOURCE: ICE, Bloomberg, and PIMCO as of 31 December 2024. Full sample based on monthly returns from 31 July 2000. Equity = S&P 500 Index; HY = ICE BofA US HY Index, Core Bond = BBG US Aggregate Index. Hypothetical example for illustrative purposes only. Exhibit is not indicative of the past or future results of any PIMCO product or strategy. There is no assurance that the stated results will be achieved. Not intended as a recommendation, nor does it represent any particular PIMCO product or strategy.
ABF proxy: We proxy monthly ABF PME returns as cash return plus excess returns over duration-matched treasuries from 75% ICE AA-BBB US ABS Index (R0A2) and 25% iBoxx US Non-Agency RMBS Index (IBXXBM38). Before February 2012, no data is available for IBXXBM38, so the spread component of the ABF Proxy is based 100% on the excess returns of ICE AA-BBB US ABS Index (R0A2). We add the return from about three years of interest rate duration exposure each month (BBG US Treasury 1-5Yr Index (LTR1TRUU) – cash). Leveraged loans: Represented by CS Leveraged Loan Index. When we leverage ABF, we keep rates exposure unchanged (about three years). We assume a time-varying financing cost above cash when leveraging the PME. To capture the credit quality differences, the financing spread for ABL is based on U.S. financial corporate debt rated A; direct lending is based on U.S. financial corporate debt rated BBB. Cash return is based on the FTSE 1M T-Bill Index.
* iBoxx US Non-Agency RMBS Index (IBXXBM38) incepted in February 2012. ABF Proxy returns include nonagency RMBS starting in February 2012.
** Volatility is measured as standard deviation of monthly returns. Sharpe ratio = (portfolio return - cash return) / portfolio volatility.
*** Value-at-Risk (VaR) is an estimate of the expected loss at a desired level of significance, based on the historical distribution of monthly returns.
Over the past three to five years – a period marked by rising interest rates and increasing correlations – the ABF proxy moderately underperforms public direct lending proxies. This underperformance can be attributed to favorable conditions for corporate credit stemming from tightening credit spreads, coupled with a negative impact on ABF performance due to interest rate exposure, despite strong underlying structured credit performance. Looking ahead, ABF may benefit from its moderate duration profile as interest rates stabilize and resume their traditional role as a diversifier against credit spreads.
However, both public securitized debt and bank loans exhibit a negative skew, indicating that their return distributions are shifted to the left.3 This serves as a good reminder to consider additional risk measures beyond volatility when constructing credit portfolios. Notably, ABF beta has a lower correlation with stocks and high yield (HY) bonds than direct lending beta, making it more positively correlated with core bonds.
Developing a forward-looking ABF model with a framework for estimated returns
Having established an attractive risk-reward profile for ABF based on historical data, we turn our attention to current valuations and future return estimates. Our approach goes beyond broad market ABS, adopting a more nuanced perspective to construct sector-specific ABF risk factors across consumer, nonconsumer, diversifying asset-backed credit, and residential mortgage debt. We analyze over two decades of monthly credit spread changes from 10 public indices4 covering major ABF asset types. These include student loans, credit cards, and auto loans on the consumer side, as well as aircraft, utilities, and equipment finance on the nonconsumer side.
We also include proxies for diversifying ABF exposures, such as music royalties, data centers, infrastructure, receivables, and franchise loans, along with residential mortgages (see appendix for details). Additionally, we exclude ABS securities rated AAA5, which constitute a large component of public structured credit markets, as their return and risk profile is generally too low for private ABF strategies.
With corporate credit spreads across investment grade (IG) and HY markets at or near historical lows, securitized markets offer marginally better valuations. To estimate returns, we adopt a risk factor-based approach similar to PIMCO’s process for determining our capital market assumptions. Given that corporate spreads, such as those for the ICE BofA US High Yield Index, are close to all-time lows, we anticipate mark-to-market losses from mean reversion over the next five years, resulting in lower credit excess returns than the yield spread over Treasuries suggests (see Exhibit 2).
In addition to carry and price impacts from mean reversion, we account for long-term average default losses per rating category in corporate and ABS markets. We also consider the cost of leverage, which we proxy with credit spreads for U.S. financials rated A or BBB for ABF and direct lending, respectively. Furthermore, the ABF model includes a term premium for three years of interest rate exposure and an incremental return based on manager skill, as well as the illiquidity and complexity premium derived from our analysis of the median private credit fund.6
Exhibit 2: Valuation and estimated returns in ABF and corporate debt
* Based on current OAS relative to 10-year history. We use PIMCO’s models for ABF factors, ICE BofA US Corporate Index for IG Corporate, and ICE BofA US High Yield Index for HY Corporate.
** Unless otherwise specified, return estimates are an average annual return over a 5-year horizon. The estimated return is based on PIMCO’s estimates for cash and duration, current spread levels of the ICE AA-BBB US ABS Index (R0A2), and the incremental return combines manager skill and liquidity / complexity premia (as observed in the median private credit fund historically – see appendix). Please refer to the appendix for additional information on estimated returns. The estimated ABS loss is based on average impairment rates of U.S. ABS rated AA, A, and BBB U.S. from 1994 to 2022. The estimated default loss in direct lending is based on long-term default and recovery statistics from Moody’s and the spread component contains an assumption of HY spread changes over the coming five years. We assume 1.25x leverage for ABF and 0.75x leverage for direct lending.
An alternative approach, which does not require assumptions about alphas or default losses, compares the credit spreads of public and private assets of similar credit quality. As we’ve previously outlined (see our Economic and Market Commentary, “Navigating Public and Private Credit Markets: Liquidity, Risk, and Return Potential”), we believe that less-liquid private investments should provide an excess premium of about 200 basis points (bps). This premium compensates investors for potential lost alpha from active management, opportunity costs associated with the inability to rebalance, and potential cash shortfall costs for unexpected liquidity needs.
Based on recent ABF deal flow with unleveraged spread of 300–450 bps7, we estimate the current incremental spread of private over public ABS markets at 240 bps 8. In direct lending, the credit spreads of recently originated deals range from 425 to 550 bps,9 which is low versus historical levels, as spreads have materially tightened over the past two years. Using leveraged loans rated B as comparison, we estimate an average private market spread pickup of 70 bps for corporate direct lending10.
We believe that the potential for 170 bps of additional private yield pickup in ABF can provide investors more compensation for illiquidity and complexity relative to corporate debt transactions. In our view, ABF presents a much larger and more diverse opportunity set, with higher barriers to entry due to differences in sourcing, origination, data, and analytics infrastructure required for thorough underwriting. As banks aim to enhance regulatory capital efficiency, we are observing a trend toward a bank partnership model with private ABF investors. This collaborative approach replaces the traditional model of direct competition for corporate debt deals.
Diversification benefits of ABF
Our detailed modelling enables us to measure correlation structures within the ABF asset class and how an aggregate ABF risk factor interacts with other risk factors, such as corporate HY. The average correlation among the consumer, nonconsumer, diversifying asset-backed, and residential mortgage ABF credit factors is only 0.55, indicating diverse risk capture and limited overlap. This high degree of risk factor diversification and the resulting reduction in volatility are key reasons for the attractive Sharpe ratio of the ABF beta.
The correlation of 0.61 between the leveraged ABF and direct lending PMEs indicates the potential for meaningful diversification between the total returns of these two asset classes.
Exhibit 3 shows a risk factor perspective where we assess the co-movement of ABF and HY credit risks in isolation of other factors like interest rates. Over the full sample, we find a correlation of 0.59 between the aggregate ABF and HY risk factors, indicating ample room for diversification benefits. While the long-term diversification of ABF to HY corporate credit is impressive, it is important to note that, in line with most factors within the credit and equity universe, correlations tend to increase during periods of stress. For instance, the ABF-HY correlation jumps to 0.83 in months where HY spreads widen by more than 50 bps.
Exhibit 3: Meaningful diversification between ABF and HY over time
Source: ICE, Bloomberg, PIMCO as of 31 December 2024. For illustrative purposes only. Figure is not indicative of the past or future results of any PIMCO product or strategy.
Note: Conditional correlation calculated over the full sample, focusing only on months when U.S. high yield spreads widened by 50 bps or more.
ABF: We evaluate the time series of credit spread (OAS) changes of various risk factors with data starting in July 2000. ABF sector-specific OAS changes are based on the public market equivalents detailed in the appendix, and overall ABF OAS change reflects an equally weighted change of spreads across consumer, nonconsumer, diversifying asset-backed credit, and residential mortgage debt. U.S. high yield: ICE BofA US High Yield Index (H0A0).
ABF is complementary to direct lending
It is widely recognized11 that the absence of mark-to-market pricing in private markets leads to a smoothing bias and can create a misleading impression of low volatility. Since alternative investments are exposed to many of the same risk factors that influence public (and liquid) financial markets, we address the downward bias in reported volatility by developing risk factor models that aim to measure “true” economic risk.
We believe that the ABF and direct lending models12 shown in Exhibit 4 allow for consistent comparisons to traditional asset classes and are valuable for optimizing multi-asset portfolios across public and private markets. The risk factor diversification within ABF is evident, while the volatility in direct lending is primarily driven by HY corporate and liquidity risk.
Exhibit 4: High degree of risk factor diversification in the ABF model
* Unless otherwise specified, return estimates are an average annual return over a 5-year horizon. For indexes and asset class models, return estimates are based on the product of risk factor exposures and projected risk factor premia. Model risk factor exposures are based on analysis of historical index data, third-party academic research and/or qualitative inputs from senior PIMCO investment professionals. Please refer to the appendix for additional information on estimated returns. Sharpe ratio calculation: (estimated return - estimated cash return) / estimated volatility. Estimated cash return = 3.32%.
** Conditional value at risk at 95% confidence over 1-year horizon. Conditional-value-at-risk (CVaR) is an estimate of the average expected loss beyond a desired level of significance. See appendix for additional information regarding volatility estimates.
Although the leveraged credit spread duration is higher in ABF than in direct lending, its volatility and equity beta are expected to be lower. This lower volatility in ABL can be attributed to the lower volatility per unit of spread duration compared with HY corporate risk factors and the correlation benefits of combining various ABF credit risk factors. An important finding for investors with existing direct lending exposures is that these two private credit strategies may perform better together. A hypothetical equally weighted mix of ABF and direct lending yields approximately 22% higher estimated risk-adjusted returns and a 12% lower equity beta than direct lending alone.
Stress tests point to ABF’s resilience and upside potential
Several fundamental factors suggest that ABF may be more resilient than direct lending during market downturns. Due to high borrowing costs and weakening covenants, some market participants anticipate declining recovery rates on defaulted loans. In contrast, many private ABF deals are backed by hard assets, benefit from diversification across a large pool of loans, typically have higher credit quality compared to non-IG corporate debt, and feature resilient, self-amortizing cash flows that are less correlated with the economic cycle.
Our stress tests indicate that ABF could experience approximately 45% less drawdown in a recessionary shock, thanks to its diversified spread factor footprint, moderate duration, and focus on noncorporate risk. However, in scenarios where both bonds and equities sell off simultaneously (e.g., stagflation), we do not expect ABF to meaningfully outperform direct lending.
Moreover, ABF has the potential to outperform direct lending during positive growth shocks, driven by its higher (leveraged) spread exposure and the currently tight starting level of corporate spreads in public and private markets. Our stress testing assumes that credit spreads will not tighten below all-time lows, which limits the potential for capital gains in corporate compared to structured credit spreads.
Conclusion: ABF offers the potential to enhance risk-adjusted returns at the total portfolio level
Based on historical data and current valuations, we believe that private ABF portfolios can generate higher returns than direct lending funds. We also estimate that ABF has lower economic volatility and is less procyclical than leveraged direct lending or equity markets.
The potential diversification benefits within the asset class, as compared to other assets, make ABF a valuable addition, not only for private market portfolios, but also for multi-asset portfolios. Strategically incorporating ABF allows investors to thoughtfully increase their private credit allocations while potentially enhancing risk-adjusted returns at the total portfolio level.13
Appendix
Data sources for the ABF credit risk factor construction
Volatility is measured as annualized standard deviation of monthly returns. We assume 6.75 years of spread duration based on 3 years of unleveraged spread duration and 1.25x leverage for ABF.
The Miscellaneous ABS category (R0O2) includes various exposures, including but not limited to music royalties, datacenters, infrastructure, wireless device payment plan agreement receivables, and whole business and franchise loans.
HY in column 6: Bloomberg US High Yield Index (H0A0).
Model risk factor exposures are based on analysis of historical index data, third-party academic research and qualitative inputs from senior PIMCO investment professionals.
Additional technical details
To ensure sufficient diversification within the subsector ABS benchmarks, we require at least 30 securities in any given month.
When no ex-AAA subsector ABS index is available, we leverage spread changes of a AAA-BBB index, based on the empirical relationship between the AA-BBB and AAA-BBB ABS indexes.
For the risk model, we assume an unleveraged spread duration of 3.0 years in ABF (equivalent to 6.75 years of spread duration when accounting for 125% leverage). We believe this is representative based on a sample of proprietary private deals. For comparison, as of 31 December 2024, the spread duration is 3.16 years for the ICE BofA AA-BBB US Asset Backed Securities Index and 3.10 years for the ICE BofA US High Yield Index.
In our ABF risk model, we adjust for liquidity and idiosyncratic risk based on our assumptions about the weighted-average life of ABF funds and the dispersion of incremental returns in the broader private debt category, as reported by Preqin over the PME benchmark. When combining the idiosyncratic risk of the four ABF sector models, instead of assuming a zero correlation, we conservatively assume a correlation structure that is identical to the correlation between the four ABF factors (average pair-wise correlation is about 0.55).
REFERENCES
Baz, Jamil, Davis, John, Han, Lloyd, and Stracke, Christian. 2022. “The Value of Smoothing.” The Journal of Portfolio Management.
Baz, Jamil, Han, Lloyd, and Loo, Marc-Antoine. 2024. “Three Models of the Liquidity Premium.” SSRN Electronic Journal. https://doi.org/10.2139/ssrn.4790409.
Block, Joern, Jang, Young Soo, Kaplan, Stephen N., and Schulze, Anna. 2023. “A Survey of Private Debt Funds.” NBER Working Paper No. 2023-10. http://www.nber.org/papers/w30868.
Pedersen, Niels, Page, Sébastien, and He, Fei. 2014. “Asset Allocation: Risk Models for Alternative Investments.” Financial Analysts Journal.