In 2019, we presented a return-based analysis framework that can be used to analyze complex fixed income funds such as PIMCO Income fund. In this updated blog, we apply a similar methodology to the fund as we did previously to evaluate the performance of the fund during the COVID market distress and see how well our model did at predicting the losses experienced by the fund.
The PIMCO Income Fund is very complex and invests in thousands of bonds and derivatives. This makes gaining insights into the underlying forces of such a rollercoaster ride very challenging. For our analysis, we will again use the advanced quantitative analysis available in Stylus Pro to parse the fund’s returns into a meaningful risk factor story.
Relative Performance During COVID Distress and 2020
PIMCO’s Income fund returned -8% in March 2020 due to the emergence of COVID as a threat to the world’s economies which sent equity markets into a sharp drawdown. The fund’s total return performance in 2020 ranked in the bottom quartile among Morningstar’s multi-sector and intermediate core-plus bond fund peer group, as depicted in the chart below.1 Outflows from the fund totaled $12.6B in March 2020 alone which, combined with the losses in the fund, decreased assets in the fund from $137B to $117B in one month. However, the fund has experienced significant recent growth in its assets as they are back to $134B as of February 2021 and Morningstar recently upgraded the fund’s Analyst Rating back to Gold.
PIMCO Income Fund’s Style
To analyze the fund, we used Dynamic Style Analysis (DSA) available in Stylus Pro because traditional rolling window regressions fail to distinguish between correlated factors, especially during periods of distress. We applied DSA to monthly returns of the fund using the same set of factors as in the last-years analysis of PIMIX. We highlight the contrasting style portfolios of the PIMCO Income fund and the Bloomberg Barclays Aggregate bond index by showing them adjacent to each other in the chart below. Different colors represent the varying exposures to fixed income factors, often referred to as risk factor betas, over time.
There are significant differences between the PIMCO Income fund and its benchmark which is widely acknowledged in the industry. The first is that the fund has historically taken more non-interest rate risk than its benchmark. The fund has historically taken very little corporate bond exposure while its benchmark has had persistent exposure to it. Conversely, the interest rate exposure taken by the PIMCO Income Fund has been on a downward trend since 2013 but slightly increased recently. The fund appears to have a persistent short position in Japanese debt.
What really sets the fund apart from its benchmark is its exposure to non-agency MBS, emerging market debt, high yield debt, asset backed debt, and CMBS. The exposures to each of these risk factors has fluctuated historically, but exposure has been continuous. But even within these factor exposures, there are marked changes over time with the exposure to non-agency MBS decreasing and to asset backed increasing since 2012.
Nominal Performance
The chart below depicts the performance of the PIMCO Income Fund, its style portfolio, and its benchmark since 2011. It shows that the PIMCO fund had a significantly larger drawdown than its benchmark during the COVID market distress, which was captured exceptionally well by its style portfolio. The fit of the style portfolio is very high, with the R-squared being 95% and MPI’s Predicted R-squared being 88%, which implies that 95% of the fund’s returns are explained by the style portfolio in sample and 88% of returns could be explained “out of sample.”
As we mentioned in the previous research, the fund’s exceptional performance through 2018 could be fully explained by its strategic factor exposures, especially non-agency RMBS. Here we also find that both the fund’s 2020 losses and spectacular recovery were also driven by the same factor model rather than dramatic shifts in the strategy. Below we will determine which risk factors contributed to this roller coaster ride.
Performance During COVID Distress
PIMCO’s income fund lost -8.4% between February and March 2020 as the extent of the COVID pandemic became apparent and global stock markets experienced significant losses. We can use our style portfolio to attribute the total return of the fund to its various factor exposures. What is quite remarkable is that the style portfolio drew down by 8.3% in comparison to the fund which drew down by 8.4%, thereby explaining largely all of the loss experienced by the fund.
The attribution analysis shows that the fund’s exposure to interest rates, which are government bonds, contributed positively to returns as investors retreated to the safety of government bonds and drove yields down. The fund’s other exposures to riskier bond factors contributed to negative returns as the spread on these riskier bonds increased due to the COVID distress. The only exception is the short Japanese bond exposure which had a non-material positive contribution to returns.
COVID Recovery
Subsequent to the market distress due to COVID in February and March of 2020, world equity markets staged a dramatic rebound as central banks acted quickly to stem the economic damage caused by COVID and stimulative fiscal policy also served to boost economies and asset markets. The PIMCO Income Fund also staged a strong recovery and returned 14.5% between April 2020 and February 2021. In the performance attribution chart below, we see the exact opposite of that observed during the COVID distress, government bonds had a negative contribution to returns while the other risk factors had positive contribution to returns as the fear of losses brought on by economic distress due to COVID decreased and spreads came in. We again see that the style portfolio (Style – blue) did an exceptionally good job of estimating the fund’s returns (Total – red) as largely all of the fund’s returns were explained by the model.
Illiquidity During COVID Distress
The COVID market distress provided a natural experiment to measure how illiquidity in the bond market, and mutual funds’ exposure to less liquid bonds, affected their performance. This topic of illiquidity in bond markets is especially pertinent for the PIMCO Income Fund as it was known for its significant non-agency RMBS holdings, which are notoriously illiquid.
In our 2020 research we demonstrated how quantitative illiquidity indicators can help investors measure liquidity risks in fixed income portfolios. The chart below shows a ranking of illiquidity based upon the funds’ aggregate illiquid asset exposure on the X axis, and the performance of each fund in March of 2020 on the Y axis. There is a notable trend between funds with more illiquidity and larger losses during March of 2020. The PIMCO Income fund performed in the middle of the range of funds with similar illiquidity.
The two dramatic outliers in the chart were the AlphaCentric Income Opportunities Fund and the Braddock Multi-Strategy Income Fund, both of which experienced losses dramatically higher than any other funds in this peer group. The fact that both of these funds are in the decile of the most illiquid funds is indicative that investors should consider quantitatively monitoring bond fund liquidity using MPI’s approach. As we noted in our previous PIMIX research, the fund’s exposure to non-agency MBS decreased significantly over the past 10 years to levels similar to the peer average. Had the fund maintained the same level of these illiquid bonds, its losses could have been of similar magnitude as the above funds.
Summary
The PIMCO Income fund had a large drawdown in the Spring of 2020 associated with the financial distress caused by the COVID pandemic. This resulted in the fund performing at the bottom of its peer group in 2020. Yet so far in 2021, it is back in the top quartile of peer funds and has been upgraded back to gold status by Morningstar. Through our returns-based quantitative analysis we discovered that most оf the fund’s ups and downs could be explained by its factor “bets” which remained relatively stable during this period. The fund suffered larger losses than its benchmark during the COVID distress because it had significantly higher exposure to non-interest rate risk and also had a strong recovery for the same reason. Being able to precisely identify these risk exposures is imperative for robust fund performance evaluation.
Footnotes
- 1Although Morningstar categorizes PIMIX as a multi-sector bond fund, for the peer group we used a combination of the two categories: multi-sector bond and intermediate core-plus, selecting single share class per fund and including only funds with at least 36 months history. Therefore, our rankings could differ from the ones published by Morningstar.