Stress Testing a Global Evergrande Shock

China Evergrande’s liquidity issues continue to stir up fears among investors of a localized downturn that could send shockwaves across global markets. Plans by Hopson Development Holdings to acquire a 51% stake in Evergrandecould provide at least some short-term wiggle room for the company to remedy its cash flow problems. Even if this proves to be a lifeline to Evergrande, there’s already evidence that the liquidity strain on Chinese real estate developers is spreading amid reports of ratings downgrades on Fantasia Holdings and Sinic Holdings due to missed debt payments.

Last week, we introduced the Axioma Worldwide 4 Linked (“Axioma WW Linked”) risk model as a tool to profile Evergrande and highlight the interconnectedness of the “Evergrande factor” (EM4 Real Estate Management & Development) with style and sector factors outside of China. Using this tool, we can dissect the correlations among the regional risk model factors and outline a stress testing framework to quantify the systematic impact of an Evergrande default on investor portfolios.

Zooming in on the “Evergrande Factor”

Before we dive into the stress testing framework, we will start by analyzing the EM4 Real Estate Management & Development (“EM REMgmt”) factor to understand its risk and correlations with other regional factors.

As shown in the chart below, the EM REMgmt factor has experienced a sharp uptick in predicted volatility within the past month, increasing from ~8% annualized volatility at the start of September 2021 to the current level of 11%.

This increase in factor volatility might be less concerning if it wasn’t for the impact of this factor on other global styles and sectors. Using the current risk model covariance matrix, we sorted the 400+ factors in the risk model in descending order based on the correlation of the EM REMgmt factor to each factor. Out of the top 5% of factors (approximately 40 factors), over 50% were style and sector factors outside of the Emerging Markets. This observation persisted for the top 10%, top 25%, and top 50% of factors based on correlation to the EM REMgmt factor.

The EM REMgmt factor's interconnectivity with global markets causes further concern due to these correlations rising in the current market environment. Let’s take the US Construction Materials factor as an example. This factor has the highest correlation to the EM REMgmt factor out of all US-style and sector factors, which makes intuitive sense given the ties between the US construction market and global real estate developers. Though this correlation has fluctuated since the beginning of 2020, it has risen sharply on a year-to-date basis, especially in the last several months.

What Strategies Will Suffer Most from an Evergrande Shock?

To build intuition on what types of strategies may suffer most from the stress tests, we can review the correlations of the style factors to the EM REMgmt factor.

Based on the above correlations, US value managers, as defined by those with high exposure to the Earnings Yield, Value, and/or Dividend Yield factors, would be hurt by a Chinese real estate drawdown. However, US growth strategies or managers with high volatility exposure could benefit from a drawdown in Chinese real estate, as illustrated by the negative correlations of these factors to the EM REMgmt factor. A similar story is true for developed markets strategies.

Stress Testing the “Evergrande Factor”

In addition to looking at the correlations between the EM REMgmt factor and the regional style factors, it's also essential to consider the total interactions across all factors, which the risk model covariance matrix can capture. To assess the systematic impact of an Evergrande shock on global markets, we'll apply the following methodology to standard market indices and a proxy portfolio from Wolfe Research that approximates the top long and short hedge fund positions in the US.

1. Estimate a shock value for the EM REMgmt factor to represent the Evergrande default scenario.
2. Multiply the EM REMgmt factor correlation to each factor by the shock value from step 1 to derive an expected factor return for each factor.
3. Multiply the expected factor return by the index or portfolio factor exposures; sum across all factors to derive the expected index or portfolio return under the Evergrande default scenario.

The stress scenario will assume a 5-standard deviation move in the EM REMgmt factor if Evergrande defaults. Furthermore, given that the recent one-month realized volatility of the EM REMgmt factor is 1.14%, the stress scenario will assume a 5.70% negative shock to the factor over the next month.

Using the above framework, we can simulate the expected return of several market indices and the US HF Proxy Portfolio under the Evergrande default scenario.

We see that some of the hardest-hit segments of the market in our scenario occur within US sectors. US Utilities, Real Estate, Consumer Staples, and Financials are sectors that could feel the most pain if Evergrande defaults. Interestingly, we see that US Hedge Fund Proxy benefits from the stress scenario, which highlights that US hedge fund managers are well-positioned to weather this impending storm.

What Happens When All Correlations Go to 1?

The expected returns above are based on the factor correlations as they exist today. But, of course, we know that during market stress, correlations tend to increase. So, while we don't know the magnitude of the rising correlations, we can look to past crisis periods to inform how this might transpire.

To do this, we re-ran the Evergrande default scenario using the risk model covariance matrix from two recent crisis environments:

1. COVID downturn, proxied using the covariance matrix from March 23, 2020
2. 2018 US-China Trade War, proxied using the covariance matrix from December 24, 2018

Under these market environments, we start to see different expectations form around the market segments most susceptible to negative returns due to an Evergrande default.

If correlations look similar to March 2020, DM and EM strategies, along with US Real Estate and Consumer Discretionary managers, will likely be the hardest hit. Conversely, if correlations are closer to those we saw in December 2018, EM strategies and US Consumer Staples, Real Estate, and Utilities managers will suffer most. Again, in both cases, US hedge fund managers appear to be defensively positioned for a market crisis.

US & Global Market Summary

US Market: 10/04/21 - 10/08/21

• All three indexes saw weekly gains. The S&P 500 and Dow advanced 0.8% and 1.2%, respectively, it was the S&P 500’s best week since August. The Nasdaq added 0.1% for the week.
• Volatility has returned to markets with both the S&P 500 and Nasdaq Composite recording back-to-back swings of at least 1% and other assets recording large moves.
• The Department of Labor said Thursday that jobless claims for the prior week totaled 326,000, lower than the 345,000 economists had been calling for.
• Lawmakers struck a deal for a short-term extension to the debt limit, stoking a rally after several days of uncertainty.
• Oil prices jumped to a seven-year high while cotton prices have been trading at their highest levels in a decade.

Normalized Factor Returns: Axioma US Equity Risk Model (AXUS4-MH)

• Earnings Yield’s 5-week freefall into oversold terrain came to an emphatic end as it backtracks just past the oversold threshhold and vaults to the top of this week’s US factor leaderboard.
• Value rose for the second straight week and pushes further into overbought territory.
• Profitability saw some positive traction but continues to be oversold.
• Size stabilized after a signficant drop last week.
• Growth slid yet again and continues its approach to the ‘Extremely Oversold’ threshold of 2.0.
• Volatility ended its 5-week tear into positive territory and finishes the week in a rare position at the bottom of the US factor leaderboard.

Despite the week’s market upswing, no single Style factor left it’s fingerprint in the returns, similar to last week’s sell-off. We also see a trivial 8 bp decrease in total US market risk.

Normalized Factor Returns: Axioma Worldwide Equity Risk Model (AXWW4-MH)

• Value continues its journey into the overbought realm and ends the week on a strong note atop the global factor leaderboard.
• Exchange Rate Sensitivity moved upward and finally broke through the overbought threshold.
• Size and Momentum continue to show weakness along with Growth which slides deeper into oversold status.

Similar to the US markets, the global markets didn’t see any dramatic changes in Style factor performance.

Regards,
Alyx