Bond markets suffered one of their worst years in decades in 2022 as central banks, including the Federal Reserve (Fed) and Reserve Bank of Australia (RBA), hiked rates aggressively in their battle to tame inflation. Given the experience of 2022 and a volatile first quarter of 2023, it is understandable that investors are wary. However, we believe that the rate hiking cycle, which contributed to recently elevated volatility across asset classes, is coming to an end in most developed markets. In Australia, as we noted in a recent article, we estimate that a 3.5%-4% cash rate in 2023 will be close to the most restrictive Australian households have experienced in terms of the percentage of their disposable income that will be required to service their debt. We think this limits the extent that the RBA can hike beyond 4% without creating material financial stability risks.

We believe that recessions are likely across developed markets in 2023-2024, although we expect them to be modest. Consistent with that macro picture, equity markets appear more vulnerable than usual. Based on PIMCO’s equity fragility model, which considers valuation, technical, and macroeconomic indicators, we estimate the probability of a large equity sell-off (correction of 20% or more within a year) at 75% in the U.S. This conditional probability is several standard deviations above the unconditional probability at 13% and is primarily driven by elevated inflation, expensive equity valuations (based on dividend yield), the (negative) momentum in one-year equity returns, and the yield curve slope^Footnote1. We estimate the equity sell-off probability, while still above long-term averages, to be lower in Australia given momentum and valuation levels paint a more favourable domestic picture.

From an asset allocation perspective, the outlook for fixed income has become much more constructive. Following central bank moves over the past year, bonds now offer more attractive yields than they have in several years. We also believe that with today’s uncertainty, fixed income can act as an important building block for defence, income, and diversification, offering the potential for both attractive returns and mitigation against downside risks, particularly in a recessionary environment.

Why yield matters: Higher return prospects for bonds going forward

In 2022, 10-year U.S. and Australian government bond yields both rose by around 2.4% as the Fed and RBA hiked cash rates by around 430 basis points (bps) and 300 bps, respectively. This recalibration to higher levels of yield may provide an attractive entry point for multi-asset investors. In a 2022 paper, “Are 60/40 Portfolio Returns Predictable?”, we showed that a simple model of expected total returns for fixed income that uses yield-to-maturity (YTM) as an unbiased predictor works particularly well for longer time horizons (such as five to 10 years). Based on data from 1951 to 2022, YTM turns out to be a significant predictor of future bond total returns.^Footnote2 A simple regression with YTM being the only explanatory variable already explains more than three quarters of the variation in returns. We think that this high predictability of bond returns can be a useful starting point for asset allocators.

In addition to the higher return prospects, a higher level of yield reduces the probability of a negative return going forward. The yield creates a cushion against potential increases in the risk free rate beyond what is currently priced into markets. With Australian 10-year yields at around 3.4% (as of 31 March), it would take a shock of +78 bps for the Bloomberg AusBond Comp 0+ Index to post a negative return over one year. At the end of 2021, with yields much closer to zero, it would have only taken a shock of +35 bps to result in a negative annual return for core bonds. Figure 1 illustrates the return profile of Australian bonds for the end of December 2021 compared with that of the end of March 2023 based on yield shocks, highlighting the greater cushion that now exists.

Further increases in yields (beyond current market expectations) could lead to short-term losses, but they would improve the reinvestment opportunity and could lead to higher returns over the long run. With expected return of around 4% for core bonds (based on YTM), we find that total returns are relatively stable around this figure for a range of macroeconomic conditions. We stress tested core bonds for shocks to GDP and inflation to simulate forward-looking one- and five-year returns as of 31 December 2022.

As Figure 2 shows, under a one-year time horizon, returns would take a modest hit in the ‘traditional growth’ scenario where inflation and GDP exceed expectations, but would be positive in all other scenarios. However, over a five-year horizon, the estimated returns demonstrate that fixed income can act as stability mechanism in a multi-asset portfolio since they are positive under all scenarios in our simulation. The reason for this is that additional increases in yield, while painful in the first year (reflected in the historical return shown for 2022 – grey bar in Figure 2), should result in higher carry being earned over the longer term.

We also looked at historical performance, analysing returns over all five Fed hiking cycles since 1988. We found that, on average, core bonds underperformed equities in the year before and during the hiking cycle. However, for the 12 months after the hiking cycle, rate sensitive sectors tend to outperform credit and equities, with the average one-year core bond return at +9.8%. Hence, investors have historically been well served by increasing exposure to fixed interest towards the end of a hiking cycle – current market pricing implies that rate hikes will end in the first half of 2023 in both the U.S. and Australia.

Credit offers some attractive opportunities for fixed income investors

Credit spreads are elevated but not pricing in a deep recession. In investment grade credit, higher all-in yield and spread levels offer investors both improved opportunities for income generation as well as a greater downside cushion, given our base case view of a mild recession. We think that the current entry point for valuations looks relatively attractive for long-term investors in investment grade and securitised credit. Our view on generic high yield corporates is more nuanced as we expect variation in sector and company fundamentals to translate into increased sectoral dispersion and ratings decompression, but this should create opportunities for active management. On securitised debt, senior non-agency mortgage-backed-securities (MBS) continue to be one of PIMCO’s highest conviction ideas while we also think that collateral loan obligations (CLOs), select asset-backed securities (ABS), and commercial mortgage-backed securities (CMBS) are trading at attractive valuations and fundamentals remain resilient.

Looking at U.S. corporate bonds across the risk spectrum, there is a high credit risk premium and expected default loss can only explain a small portion of the credit spread (see Figure 3). For example, the yield spread for BBB corporates is close to 100 bps above the expected losses from default even with defaults rising to the 95^th percentile.

Credit also looks cheap relative to equities given that debt is a senior claim but is currently offering a similar yield to equities. As of 31 December 2022, the expected return on the S&P 500 (calculated from the cyclically-adjusted earnings yield plus 10-year break-even inflation rate) and the default-adjusted yield on 10-year maturity BBB credit are 5.9% and 5.7%, respectively. The average difference between these two measures of forward-looking return is larger than that and the current spread is in the 29^th percentile based on the last 100 years of data. This implies that credit is currently cheap relative to equities. A similar phenomenon can be observed in Australia as yields on many high quality corporate bonds are higher than the dividend yield of the corresponding stock.

The stock-bond correlation should remain negative in recessionary periods

For the past 20 years, the stock-bond correlation has, on average, been negative. However, in 2022 the correlation was positive at +0.38 and +0.44 in the U.S. and Australia, respectively – representing extreme values above the 95th percentile.^Footnote3 As we discussed in a 2019 paper, “Stocks, Bonds and Causality”, this flip in the correlation coefficient is hardly surprising given that periods of elevated bond volatility were historically associated with a positive correlation between stocks and bonds. Generally, we find that shocks to the equity market tend to be associated with a negative relationship, while shocks to real bond yields and expected inflation are associated with a positive relationship between stock and bond returns.

It is important to note that regardless of the longer-term correlation, bonds have typically provided positive returns and outperformed equities in recessionary periods. In fact, excess bond returns have been positive in nine out of the last 10 recessions in the U.S., and we expect this diversification property to remain intact in the event of a material economic growth shock. With yields much higher than a year ago, the room for a potential bond market rally in a recessionary shock is much higher.

If investors believe that central banks remain credible and longer-term inflation expectations remain anchored while realised inflation gradually returns to target, the stock-bond correlation is likely to revert towards longer-term averages of around zero. This is consistent with our view that central bank actions will be more muted in 2023, and the market’s focus may be shifting from inflation toward concerns about economic growth.

Diversification benefits of fixed income in an equity-dominated portfolio remain intact

Most Australian portfolios have a large allocation to pro-cyclical risk. While institutional investors take embedded equity beta through allocations to illiquid assets like property, infrastructure, and private equity, many retail investors hold a significant quantity of Australian bank hybrid securities. This overweight to risk assets may have appeared reasonable when government bond yields were close to all-time lows, but the optimal stock-bond portfolio mix has changed materially given the repricing seen in 2022.

We analysed the risk and return characteristics of core bonds, liquid multi-sector credit, and private credit as well as a combination of these three assets.^Footnote4 The diversified fixed income portfolio demonstrated an attractive expected return not far below equities with less than half of the risk (see Figure 4). Therefore, the expected Sharpe ratio of a diversified fixed income portfolio is materially higher than that of the diversified equity benchmark.

Our analysis of the drivers of risk in each portfolio shows that lower volatility is achieved through the correlation benefits of combining multiple risk factors (see Figure 5). At the diversified fixed income portfolio level, interest rate duration remains the dominant risk factor, but risk exposures to credit, idiosyncratic risk, and illiquidity provide diversification.

After considering fixed income in isolation, we looked at the efficient frontier^Footnote5 to assess the optimal combination of equity and fixed income assets. The efficient frontier allows any combination of fixed income assets to be added to domestic and global stocks. The pure stock portfolio^Footnote6 is located below the efficient frontier with an unfavourable risk-return trade-off, indicating that adding fixed income provides material efficiency gains (see Figure 6). The chart also indicates the improvement in the Dec-2022 efficient frontier relative to the same investable asset universe in December 2021.

Incorporating fixed income into an equity-dominated portfolio has a positive impact on risk-adjusted expected returns (as measured by the Sharpe ratio). Figure 7 shows that the Sharpe ratio of an equity-dominated multi-asset portfolio is expected to increase with a growing allocation to the diversified fixed income portfolio.

Figure 7 also shows the result of a portfolio stress test. The top line (long-term correlations) uses estimated correlations between all equity and fixed income assets calculated from risk factor returns since 1997. From this base case, we introduce constraints on the forward-looking correlations between all assets by enforcing a positive correlation between all assets. The lowest line shows the Sharpe ratios when the correlation coefficient between all assets is constrained to be at least +0.5. The key takeaway is that even with strong positive correlation between stocks and bonds, the fixed income component adds meaningful Sharpe ratio improvements to an equity-dominated portfolio and a balanced portfolio of stocks and bonds remains optimal. The benefit of diversification, even if investors are wrong about forecasting the correlation, remains intact as long as the assets are not perfectly correlated.

Bonds continue to play a key role in a diversified portfolio

From an asset allocation perspective, the outlook for fixed income has become much more constructive in recent months. Investors with long time horizons have the opportunity to lock in high levels of yield and benefit from the increased diversification potential in fixed income

The fixed income universe is broad and ranges from assets that are low to negatively correlated to equities, to those that are highly correlated with stocks. This means diversification and sensible portfolio construction are key to meeting investor objectives.

While core bonds should remain the anchor to a traditional fixed income portfolio, complementary strategies like multi-sector credit, or private credit strategies may aide performance during periods of elevated interest rate volatility.

Our analysis shows that adding diversified fixed income exposure to multi-asset portfolios that are dominated by equity risk materially improves the expected Sharpe ratio even if the stock-bond correlation were to be strongly positive.