Modeling Securitized Credit

Opportunity: Exposure to a Diverse Set of Risks

The evolution in securitized credit markets following the financial crisis has led to a broad array of opportunities spanning sectors such as residential housing, U.S. consumers, corporate credit and commercial real estate. Securitized credit encompasses investment grade, below investment grade and nonrated securities as well as fixed and floating rate coupons.

Challenge: Securitized Credit is Difficult to Index

At the same time, the fragmented nature of the securitized credit universe has challenged index providers. The confluence of these events has “broken the mold” for index creation, as many indices are either investment grade or high yield, fixed rate or floating rate, while other taxable fixed income indices require a bond to have a rating for inclusion. As a result, the few indices that exist only cover a portion of the market; for example, the securitized portion of the Bloomberg Barclays Aggregate index is agency dominated, JPMorgan’s CLOIE index covers collateralized loan obligations (CLOs) and the Vista CRT indices cover credit risk transfer securities. Investors and asset allocators are limited in how they can model and identify appropriate strategic allocations to this important part of the fixed income market. Without a reasonably indicative, reliable data set for use in a portfolio optimization, how can investors be confident in their allocation decisions? Consequently, investors have systematically underallocated to this opportunity set.

Solution: Build a Strategic Allocation Model

To address this challenge and help investors understand how to evaluate and allocate to this important asset class, we conducted a detailed analysis of historical comparative returns from January 1, 2013 through March 31, 2020. To better understand the impact from the recent market dislocation, we separated our analysis into pre- and post-Covid-19 periods. The step-by-step process to arrive at an asset allocation decision to this opportunity-rich sector is described below.

Key Analytic Takeaways

  • Securitized credit has historically enhanced the risk-adjusted returns for a 50/50 senior loan and high yield portfolio (pre- and post-Covid-19)
  • The improvement in Sharpe ratios for both data sets was primarily driven by a decline in annualized volatility, highlighting the diversification benefits of adding a securitized credit allocation to portfolios
  • The newness of this research demands multiple perspectives and optimizations, leading to a range rather than a pinpoint recommendation: Our analysis shows that a band of 34–48% represents an appropriate range for allocations to securitized credit
  • We believe the consistency of the pre- and post-tail event data sets validates our belief that investors should include a strategic allocation to securitized credit

Constructing an Analytic Framework

Step 1: Define securitized credit

Step 2: Identify securitized credit constituents

Step 3: Build a performance series

Step 4: Define “home base,” i.e., optimizing to what? Step 5: Determine strategic allocation bands

 

Step 1: Define Securitized Credit

Before we can optimize for securitized credit allocations, we need to define what comprises the space. Securitized credit spans a diverse range of sub-sectors, driven by different cycles with a range of credit ratings, offering fixed and floating rate coupons. The investment landscape includes asset-backed securities (ABS), collateralized loan obligations (CLOs), commercial mortgage-backed securities (CMBS), non-agency residential mortgage-backed securities (RMBS), credit risk transfer securities (CRTs) and traditional agency RMBS. Figure 1 highlights key attributes of these asset classes.

Figure 1. Securitized credit represents many flavors of risk
Figure 1. Securitized credit represents many flavors of risk

Source: Voya Investment Management as of March 31, 2020

The building blocks for securitized credit, with the diversity of choices and investment opportunities, is accompanied by challenges and complexity for asset allocation purposes. With no single index or basket of indices properly capturing this market beta, we elected to build one.

 

Step 2: Identify Securitized Credit Constituents

To identify securitized credit constituents, three key characteristics underpinned our framework:

  1. A dedicated securitized strategy that
  2. Invests in multiple subsectors within the securitized credit universe and
  3. Has institutional and individual visibility and availability

Figure 2 summarizes the framework and the number of constituents after each step.

Figure 2. Voya framework for identifying securitized credit constituents
Figure 2. Voya framework for identifying securitized credit constituents

Source: Morningstar Direct, eVestment and Voya Investment Management as of March 31, 2020

What emerges after our analysis is a list of ten securitized credit constituents for the creation of a performance series to use for asset allocation purposes. Figure 3 summarizes these constituents as they appear in Morningstar Direct as well as eVestment. While we acknowledge the list may have overlooked some potential candidates for inclusion, we believe it is robust enough to move forward with our research agenda. In addition, as peers across the industry conduct similar research and perhaps expand the list of the securitized credit constituents, we believe this will merely fortify the case for securitized credit and further validate our research conclusions.

Figure 3. Securitized credit performance series comprises ten constituents
Figure 3. Securitized credit performance series comprises ten constituents

Source: Morningstar Direct, eVestment and Voya Investment Management as of March 31. 2020.

*Mutual fund listed in eVestment.

Step 3: Build a Performance Series

With our constituents identified, the performance series and the underlying securitized credit portfolio were built using the following portfolio construction rules:

  1. Starting the series required that three mutual funds exist
  2. Implement equal-weighted allocations
  3. Rebalance the portfolio at the end of each calendar year
  4. Add new constituents at the end of every calendar year

As a result, Voya’s securitized credit performance series and the accompanying optimization work begins January 1, 2013. A summary of the model portfolio year-by-year constituents through 2020 is summarized in Figure 4.

Figure 4. Securitized credit model allocations over time
Figure 4. Securitized credit model allocations over time

Source: Morningstar Direct and Voya Investment Management as of March 31, 2020

Step 4: Define Home Base

With the securitized credit model portfolio built and a performance series created, the next step in this project was to define home base and understand what we would be optimizing to. For our research, we looked to identify potential funding sources for securitized credit allocations by analyzing correlations, annualized returns and volatilities with an objective of enhancing risk-adjusted returns.

In the midst of this project, Covid-19 gripped the financial markets. This event was both disruptive and instructive, presenting a unique opportunity to conduct research including and excluding a “black swan” tail event. The stability and variability in the statistics as well as the optimization results would be informative and influential. In fact, these events influenced our decision to identify allocation bands rather than to specify a pinpoint forecast.

We began with a review of correlations for the two periods as illustrated in Figure 5. We constructed correlation matrices for the periods 2013 through 2019, which excludes the Covid-19 tail event and 2013 through March 2020, to include Covid-19. Analyses of correlations and the shifts in correlations serve as a guide to which other sectors of the broad fixed income market could serve as sources for securitized credit allocations, as well as for our home base portfolio.

Figure 5. Before Covid-19, securitized credit demonstrated low correlation to high yield and bank loans — after Covid-19, correlations increased measurably
Figure 5. Before Covid-19, securitized credit demonstrated low correlation to high yield and bank loans — after Covid-19, correlations increased measurably

Source: Morningstar Direct, Bloomberg Barclays, S&P/LSTA and Voya Investment Management as of March 31, 2020. Asset class definitions: Aggregate = Bloomberg Barclays U.S. Aggregate Bond index, bank loans = S&P/LSTA Leveraged Loan index, high yield = Bloomberg Barclays U.S. High Yield 2% Issuer Cap index, MBS = Bloomberg Barclays U.S. Mortgage-Backed Securities index, securitized credit = securitized credit performance series. One cannot directly invest in an index. Past performance is no guarantee of future results.

As summarized in Figure 5, high yield and bank loans are loosely correlated to our securitized credit performance series for the preCovid-19 data set, while the Bloomberg Barclays Aggregate index and agency mortgage-backed securities had a lower correlation. Meanwhile, the correlations estimated with the post-Covid-19 data set for high yield and bank loans increased while the correlations to the Aggregate and agency mortgage-backed securities decreased. In addition to correlations, we reviewed annualized returns and standard deviations for both periods, with the results illustrated in Figure 6. Evaluating a risk/return graph for both periods crystallizes funding sources that will serve as home base.

Figure 6. Securitized credit’s risk/reward profile has held up through Covid-19
Figure 6. Securitized credit’s risk/reward profile has held up through Covid-19

Source: Morningstar Direct and Voya Investment Management. Asset class definitions: Aggregate = Bloomberg Barclays U.S. Aggregate Bond index, bank loans = S&P/LSTA Leveraged Loan index, high yield = Bloomberg Barclays U.S. High Yield 2% Issuer Cap index, MBS = Bloomberg Barclays U.S. Mortgage-Backed Securities index, securitized credit = securitized credit performance series. One cannot directly invest in an index. Past performance is no guarantee of future results.

When reviewing the risk/return plots for the two data sets, agency mortgage-backed securities (MBS) and the Aggregate hovered in a similar range for both periods. This contrasts with high yield, bank loans and securitized credit, which saw substantive shifts. Our review of risk/return in conjunction with our correlation analysis led us to a framework focused on identifying optimal allocation bands for securitized credit in the context of a home base portfolio that included high yield and bank loans. We believe this decision is intuitive and highlights clear parallels between the two asset classes. Securitized credit is primarily spread risk, higher yielding and includes fixed and floating rate instruments. In the corporate credit arena, high yield is a fixed rate opportunity set and bank loans are a floating rate opportunity set; both are primarily spread risk assets and have historically offered higher returns albeit with higher volatility.

Our home base portfolio is an equal-weighted portfolio of high yield and bank loans. Now, this is not based on naïve analysis, but recognizes this 50/50 construct has been adopted as a standard benchmark for institutional multisector credit mandates. It is important to highlight the differences between multisector credit and multisector bond strategies. For investors familiar with the Morningstar multisector bond category, multisector bonds are typically benchmarked to the Aggregate or Bloomberg Barclays Universal index. While they will invest in a significant number of securities rated below investment grade, they will also have a healthy allocation to investment grade bonds and even U.S. Treasurys. Meanwhile, multisector credit mandates are multisector, as the name suggests, with a focus on higher yielding credit investments. The investment mandates can include high yield and bank loans as well as other higher yielding fixed income sectors, including securitized credit and emerging markets.

 

Step 5: Determine Strategic Allocation Bands

Our final step includes an outline of our optimization framework, followed by a review of our constrained and unconstrained optimizations. Our optimization was constructed as follows:

  1. We focused on risk-adjusted returns, so our optimization is structured to maximize Sharpe ratios.
  2. We used a simplified framework of just three investments — high yield, bank loans and securitized credit — to better understand the relationships. It is important to note that embedded within this optimization, we incorporated another level of consideration. Voya’s securitized credit series is net of fees, while the index information we use for high yield and bank loans excludes any management fees or other expenses. We believe our deliberate decision to “haircut” the securitized credit performance series, with net of fees results, adds another threshold for the optimizer to allocate to this opportunity set.
  3. We conducted multiple optimizations to identify optimal allocation bands, as opposed to a specific pinpoint recommendation. We used the tail event that occurred in the first quarter of 2020 as an opportunity to analyze the data pre- and post-tail event. We also defined constrained and unconstrained optimizations. Our constrained optimization required equal allocations to high yield and bank loans. By contrast, our unconstrained optimization allowed the allocations to high yield, bank loans and securitized credit to vary.

Figures 7 and 8 summarize the optimization results. Figure 7 shows the home base model portfolio (50% high yield and 50% bank loans) and the results for the constrained optimizations that maintained equal allocations between high yield and bank loans while maximizing Sharpe ratio for our optimal portfolio. The constrained optimization recommended identical allocations for both data sets. For both pre- or post-Covid-19 data sets, the optimal portfolio included an allocation of 29.5% to high yield, 29.5% to bank loans and 41.0% to securitized credit. We believe the consistency of the pre- and post-tail event data sets validates our long-held belief that investors should have a strategic allocation to securitized credit. Beneath the allocations, Figure 7 summarizes annualized returns, standard deviation and Sharpe ratio for the home base portfolio and constrained optimizations for pre-Covid-19 on the left and post-Covid-19 on the right. The improvement in Sharpe ratio for both data sets is primarily driven by a decline in annualized volatility, highlighting the diversification benefits of a securitized credit allocation in portfolios.

Figure 7. Voya constrained optimization results
Figure 7. Voya constrained optimization results

Source: Morningstar and Voya Investment Management. Pre-Covid-19 covering the period January 1, 2013 – December 31, 2019. Post-Covid-19 covering the period January 1, 2013 – March 31, 2020. Asset classes are represented as defined in previous figures. One cannot invest directly in an index. Past performance is no guarantee of future results.

Figure 8 summarizes the output for the unconstrained optimizations, including information for the pre- and post-Covid-19 data sets. The unconstrained optimization results show a meaningful shift in optimal portfolio allocations, quite unlike the constrained optimizations. Whereas the unconstrained optimization using the pre-Covid-19 data set has an allocation of 17% high yield, 49% bank loans and 34% securitized credit, the optimization using the post-Covid-19 data set through March 31, 2020 shows a very different portfolio with an allocation of 52% to high yield and 48% to securitized credit. Annualized returns, standard deviations and Sharpe ratios for the pre- and post-Covid-19 data sets are also summarized in the display.

Figure 8. Voya unconstrained securitized credit optimization results
Figure 8. Voya unconstrained securitized credit optimization results

Source: Morningstar and Voya Investment Management. Pre-Covid-19 covering the period January 1, 2013 – December 31, 2019. Post-Covid-19 covering the period January 1, 2013 – March 31, 2020. Asset classes are represented as defined in previous figures. One cannot invest directly in an index. Past performance is no guarantee of future results.

The shifts in allocations to securitized credit highlight why we frame the recommendation in a band as opposed to a pinpoint target. We also believe it is important to shed some light on why the weights changed meaningfully, including the abandonment of bank loans in the unconstrained optimization using the post-Covid-19 data set.

We believe analyzing a risk/return radar chart offers clues on the unconstrained optimization behavior. Figure 9 plots the annualized return and annualized volatility for securitized credit along the axis going to the top, bank loans in the lower right and high yield in the lower left. The light blue radar triangle represents the plotting of the annualized return and the dark blue radar triangle represents the plotting of the annualized standard deviation for the pre- and post-Covid-19 data sets.

Figure 9. Risk/return radars offer clues on optimization behavior
Figure 9. Risk/return radars offer clues on optimization behavior

Source: Bloomberg Barclays, Morningstar Direct, S&P/LSTA Leveraged Loan Index and Voya Investment Management. Modeling securitized credit performance series based upon an equal-weighted portfolio of securitized credit mutual funds, rebalanced annually. Constituent asset classes and years of inclusion as defined in previous figures. One cannot invest directly in an index. Past performance is no guarantee of future results.

In the pre-Covid-19 data, there is a distinct leaning towards high yield, illustrating its higher annualized returns and higher standard deviation. There is decidedly less volatility in securitized credit and bank loans, albeit with a modest give-up in return. With our optimization focused on maximizing Sharpe ratios, it makes sense that the allocations gravitated to bank loans and securitized credit with a modest allocation to high yield. We compare that to the risk/return radar chart for the period of January 2013 through March 2020, which represent a post-Covid-19 tail event data set. There are two key takeaways. First, the radar triangles have flipped. With the steep and rapid sell-off in 1Q20, the volatility radar now spans more area than the annualized return radar in the right-hand side of Figure 9. Second, the volatility radar on the right is more equal (closer to an equilateral triangle) than the volatility triangle on the left. With similar volatility, the optimizer moves away from bank loans and is more biased towards high yield and securitized credit, which potentially offer higher returns with a similar volatility profile.

Conclusion

We believe the research framework outlined in this paper makes a strong case for strategic allocations to securitized credit in client portfolios, likely well in excess of their existing exposure, if any. The results of our four portfolio optimizations suggest that a band of 34–48% weightings represents an optimal range for securitized credit allocations. Like other sectors and sub-sectors across the global fixed income market, the diverse nature of, and continued evolution within, the securitized credit universe present attractive investment opportunities to enhance the potential for portfolio diversification and risk-adjusted returns. Perhaps most powerfully, this methodology can now also be validated against the market’s most recent black swan event.

IM1217835

Voya Investment Management has prepared this commentary for informational purposes. Nothing contained herein should be construed as (i) an offer to sell or solicitation of an offer to buy any security or (ii) a recommendation as to the advisability of investing in, purchasing or selling any security. Any opinions expressed herein reflect our judgment and are subject to change. Certain of the statements contained herein are statements of future expectations and other forward-looking statements that are based on management’s current views and assumptions, and involve known and unknown risks and uncertainties that could cause actual results, performance or events to differ materially from those expressed or implied in such statements. Actual results, performance or events may differ materially from those in such statements due to, without limitation, (1) general economic conditions, (2) performance of financial markets, (3) interest rate levels, (4) increasing levels of loan defaults, (5) changes in laws and regulations, and (6) changes in the policies of governments and/or regulatory authorities. Past performance is no guarantee of future returns.

The opinions, views and information expressed in this commentary regarding holdings are subject to change without notice. The information provided regarding holdings is not a recommendation to buy or sell any security. Strategy holdings are fluid and are subject to daily change based on market conditions and other factors.