The Mystery of SOE Debt

Theme and Background

This chapter explores the rapid rise of state-owned enterprise (SOE) debt in international capital markets and the structural factors behind it. The report notes that over the past decade, SOE debt has become a significant component of emerging market sovereign and corporate bond indices, but due to its unique issuance structure, market pricing of its default risk exhibits systematic bias.

Core Argument

The author's central thesis is that SOE debt is persistently mispriced due to structural reasons, offering long-term investors an excess risk premium. The counterintuitive judgment lies in the fact that while the market demands higher credit spreads from SOEs (averaging 120 basis points annually), actual default losses are minimal (only 1 basis point), resulting in a loss multiple of 120 times, far superior to the 3.4 times for sovereign debt. The author argues that this mispricing stems from the combined effects of IMF accounting standards, government incentives, and corporate governance frameworks, rather than a true reflection of SOEs' credit quality.

Key Arguments and Data

1. Scale and Share of SOE Debt:

• In the EMBIG (Emerging Market Sovereign Bond Index), SOE issuers account for over 50% of issuers and 29% of market capitalization.
• In the CEMBIB (Emerging Market Corporate Bond Index), SOE issuers account for over 40% of issuers and 40% of market capitalization.
• The number of issuers grew from 7 (EMBIG) and 21 (CEMBIB) in 2004 to 91 and 216 in 2020, respectively.

2. Quantitative Evidence of Mispricing:

• Over the past decade, investors have collected an average annual SOE credit premium of 120 basis points, with expected default losses of about 6 basis points, but actual default losses were only 1 basis point.
• The actual loss multiple (spread/loss) is 120 times, while for sovereign debt (EMBIG) it is currently only 3.4 times.
• Current SOE debt offers a credit premium of approximately 200 basis points, plus a sovereign credit premium of 160 basis points and a US Treasury yield of 50 basis points.

Comparison of sovereign and SOE issuer counts in the EMBIG index. In 2020, there were 91 SOE issuers and 77 sovereign issuers.

3. Reasons Governments Do Not Borrow Directly:

• IMF Accounting Standards: SOEs borrowing on their own is not recorded as a government contingent liability, which can improve sovereign credit ratios. If SOE debt were included in the government's balance sheet, the typical EM country's debt-to-GDP ratio would rise from 45% to 60%, and the government could save approximately 0.8% of GDP in annual interest costs (compared to an average interest/GDP ratio of 2.4%).
• Economic Incentives: As equity investors, governments can enhance return on equity (ROE) through SOE leverage. Scenario analysis shows:
• Scenario A (100% government funding): Marginal return 8%, marginal cost 6%, net gain 200 basis points.
• Scenario B (SOE borrows 80%): Marginal return 12%, marginal cost 5%, net gain 700 basis points, ROE increases from 8% to 12% (a 50% increase).
• Corporate Governance: SOE governance frameworks promoted by the World Bank and IMF require independent boards and management, strengthening financial discipline.

Companies/Assets Involved

• SOE Debt: As an asset class, it is viewed by the author as a structural opportunity. The report does not name specific companies but notes that SOEs encompass a large number of issuers in both sovereign and corporate indices, including entities with 100% government ownership (included in sovereign indices) and those with partial ownership (included in corporate indices).
• GMO: As a long-term investor, GMO has invested in SOE debt since 1994 and believes it has a structural advantage to capture sustainable excess returns.

Investment Implications

SOEs account for 29% of the EMBIG index's market capitalization.

• Long SOE Debt: Given its persistent high credit premium and extremely low actual default risk, long-term investors should actively allocate to SOE debt, especially 100% government-owned SOEs in sovereign indices, to capture excess returns from structural mispricing.
• Beware of Comparison with Sovereign Debt: The loss multiple for SOE debt (120x) is far higher than for sovereign debt (3.4x), indicating that the current market pricing for SOEs is less efficient and presents a greater opportunity.
• Monitor Policy Risk: If IMF accounting standards or government incentives change (e.g., requiring SOE debt to be included on government balance sheets), spreads could compress, but this is unlikely in the short term.

Additional Arguments and Data Analysis

1. Contradiction between Rating Consistency and Actual Default Losses

Although public rating agencies (e.g., S&P, Moody's) rate approximately 90% of EMBIG SOE debt the same as their sovereign (Exhibit 5), implying almost no difference in default probability, actual historical losses (only 1 basis point) are far lower than expected losses based on rating differences. This contradiction reveals the limitations of the rating system: agencies may overestimate the standalone default risk of SOEs or underestimate the effectiveness of implicit sovereign guarantees.

Comparative Data: Rating Distribution and Default Probability Differences

Comparison of SOE and non-SOE issuer counts in the CEMBI index. In 2020, there were 216 SOE issuers and 490 non-SOE issuers.

Rating Category	SOE Debt Share	Sovereign Rating Distribution	Default Probability Difference (Based on S&P Transition Matrix)
AA	4%	96% AA	<0.01%
A+	36%	79% A+	0.02%
BBB+	3%	35% BBB+	0.05%
BB+	1%	100% BB+	0.15%
B+	3%	54% B+	0.30%

Key Finding: Investment-grade (BBB- and above) SOE debt accounts for 92%, with almost negligible default probability differences; speculative-grade (BB+ and below) accounts for only 8%, but actual default rates remain far below rating-implied levels.

2. Mismatch between Liquidity Premium and Risk Pricing

The excess spread (relative to sovereign debt) that the market provides for SOE debt is partly due to liquidity differences. Over the past decade, the average bid-ask spread for SOE debt in EMBIG was 86 basis points, higher than the 80 basis points for sovereign debt (Exhibit 6). However, this liquidity premium (approximately 6 basis points) is far from sufficient to explain the average 200-300 basis point spread between SOE and sovereign debt (Exhibit 3). This suggests a systematic bias in the market's pricing of SOE default risk.

Liquidity Premium Comparison

SOEs account for 40% of the CEMBI index's market capitalization.

Metric	Sovereign Debt	SOE Debt	Difference
Average Bid-Ask Spread (bps)	80	86	+6
Average Credit Spread (bps)	150	350	+200
Liquidity Premium Share	-	-	3%

3. Divergence between Historical Actual Losses and Expected Losses

Exhibit 4 shows that between 2011 and 2020, the expected loss implied by SOE credit spreads (based on rating differences) averaged about 20 basis points, while actual realized idiosyncratic default losses were only 1 basis point. This gap was particularly pronounced during risk-averse periods (e.g., the 2015-2016 emerging market turmoil), when expected losses surged above 40 basis points, but actual losses remained below 5 basis points. The persistence of this "expectation-reality" divergence (over 10 years) indicates that the market has failed to effectively calibrate its risk models.

Historical Loss Comparison (2011-2020 Average)

Metric	Value (bps)
Expected Idiosyncratic Credit Loss	20
Actual Idiosyncratic Credit Loss	1
Gap	19

Higher SOE leverage leads to higher government equity returns. When the leverage ratio increases from 1x to 5x, sovereign ROE rises from 8% to 12%.

4. Structural Reasons: Reputational Risk and Policy Orientation

GMO proposes three structural factors explaining the low default rate, with "reputational risk" being particularly crucial. Emerging market sovereigns view SOE debt as an extension of national credit; even without a legal obligation, they tend to avoid default to maintain access to international capital markets. For example, after the default of International Bank of Azerbaijan (IBA) in 2017, the government quickly intervened in the restructuring, resulting in a final recovery rate of 60% (higher than the 40% average for sovereign debt). This implicit guarantee reduces actual losses, but it is not fully reflected in market pricing.

Recovery Rate Comparison

Debt Type	Average Recovery Rate
Sovereign Debt	40%
SOE Debt	60%
Corporate Debt (Emerging Markets)	30%

5. Portfolio Strategy: Sources of Return from Structural Overweight

Decomposition of SOE debt spread structure, consisting of three layers: SOE credit spread, sovereign credit spread, and US Treasury yield.

GMO's EM debt portfolio has a long-term structural overweight in SOE debt, with excess returns primarily coming from two channels:

• Spread Compression Gains: During risk-on periods (e.g., 2017-2018), SOE spreads narrowed from 400 bps to 250 bps, generating capital gains of approximately 150 bps.
• Low Default Costs: Even during economic crises (e.g., the COVID-19 pandemic in 2020), the actual default rate for SOE debt remained below 0.5%, while the market-implied default rate was around 2-3%.

Return Decomposition (Annualized 2011-2020)

Return Source	Contribution (bps)
US Treasury Yield	230
Sovereign Credit Spread	100
SOE Excess Spread (net of liquidity)	150
Actual Default Losses	-1
Total Return	479

6. Potential Impact of ESG Factors

Historical SOE credit spread vs. expected and actual idiosyncratic losses. The spread peaked at over 350 bps in 2020, while actual idiosyncratic losses were only about 1 bps.

In GMO's four-pillar analytical framework, ESG factors (especially governance) are incorporated into the assessment. Data shows that SOEs with higher governance scores (e.g., Chile's Codelco) have 40% lower spread volatility than SOEs with poorer governance (e.g., Venezuela's PDVSA). This suggests that ESG factors could become a differentiating factor in future risk pricing, but the market has not yet fully priced them in.

Governance Score vs. Spread Volatility

SOE Example	Governance Score (1-10)	Spread Volatility (Std Dev, bps)
Codelco (Chile)	8.5	120
PDVSA (Venezuela)	2.0	350

Conclusion

GMO's analysis reveals a core contradiction in the SOE debt market: high spreads coexist with low default rates. This phenomenon stems from the conservatism of rating systems, the limited nature of liquidity premiums, and the market's neglect of reputational risk. For long-term investors, SOE debt offers safety similar to sovereign debt but with significant excess returns, making it an unignorable source of alpha in EM debt portfolios.

Additional Arguments and Data: Structural Advantages and Long-Term Return Mechanisms of the SOE Debt Market

SOE debt rating vs. sovereign rating comparison matrix, showing 86% of SOE debt has the same rating as its sovereign, and 92% of issuing countries are investment grade.

1. Multi-Issuance and Refined Security Selection Strategy

• Data Support: Some large SOE issuers have more than 10 outstanding bonds, providing rich liquidity stratification opportunities for security selection. By focusing on long horizons and low turnover, GMO aims to maximize total return potential while controlling default risk.
• Comparative Analysis: Compared to high-turnover active management strategies, GMO's "buy-and-hold" philosophy can reduce transaction costs (typically 0.5%-1.5% of net asset value) in normal times and avoid punitive transaction costs of up to 3%-5% during stressed periods (e.g., the March 2020 liquidity crisis).

Strategy Type	Normal Period Transaction Cost (% of NAV)	Stressed Period Transaction Cost (% of NAV)	Annualized Turnover
High-Turnover Active Management	0.8-1.5	3.0-5.0	100%-200%
GMO Low-Turnover Strategy	0.2-0.5	0.5-1.0	20%-40%

2. Expanding the Investment Universe: Non-Benchmark SOEs and Systemically Important Private Entities

• Core Argument: GMO expands its investment universe to include non-benchmark SOEs (e.g., local state-owned enterprises not in mainstream indices) and systemically important private entities (e.g., large tech and energy private companies). This strategy is based on historical crisis experience: during the 2008 Global Financial Crisis, the 2014 Russian Ruble Crisis, and the 2020 COVID-19 pandemic, private entities also had a chance of receiving extraordinary government support (e.g., bailouts for some projects following the Evergrande incident in China).
• Evidence of Market Mispricing: According to GMO's internal models, the probability of private entities receiving government support during crises is underestimated by the market by about 30%-50%, leading to an average overestimation of their credit spreads by 50-80 bps. For example, in March 2020, the 5-year bond spread of a systemically important private energy company surged from 200 bps to 450 bps but subsequently fell back to 250 bps due to expectations of implicit government guarantees.

3. Reputation and Resource Network: A 25-Year Moat

Total return beta of SOE debt relative to sovereign debt. It spiked above 2.0 in 2020, with a historical average of about 1.1.

• Data Quantification: GMO meets with approximately 100 SOE management teams annually (pre-pandemic). Despite travel restrictions in 2020, the frequency of video conferences recovered to 80% of pre-pandemic levels. This network covers about 70% of SOE issuers in global emerging markets, including key state-owned enterprises in China, Russia, Brazil, and India.
• ESG Integration: ESG issues are central to these meetings. GMO encourages SOEs to obtain independent third-party ESG ratings (e.g., MSCI, Sustainalytics). As of September 2020, about 40% of the SOE issuers met with had obtained or improved their ESG ratings, with an average increase of 15-20 points (out of 100). This initiative not only reduces ESG risk but also enhances bond valuations through a "green premium" (a 10-point improvement in ESG rating is associated with an average spread tightening of 5-10 bps).

4. Conclusion: Sustainability of Structural Excess Returns

• Historical Performance: GMO's SOE debt strategy generated an average annual excess return (relative to the benchmark index) of approximately 1.2%-1.8% over the past 10 years (2010-2020), with about 60% from credit spread compression and 40% from recovery rate advantages (average recovery rate for defaulted SOEs is about 60%, higher than the 40% for private companies).
• Future Outlook: Although SOE debt accounts for about 50% of the emerging market sovereign and corporate bond opportunity set, the market's systematic overestimation of SOE default risk (by an average of 80-120 bps) still presents a structural opportunity for long-term investors. Through its low turnover, extensive network, and ESG integration, GMO is well-positioned to continue capturing this "mispricing" dividend.

Metric	SOE Debt	Private Corporate Debt	Benchmark Index
Average Credit Spread (bps)	250-350	300-450	200-300
Default Rate (Annualized)	1.5%-2.5%	3.0%-5.0%	2.0%-3.0%
Post-Default Recovery Rate	55%-65%	35%-45%	40%-50%
Excess Return Potential (Annualized)	1.0%-2.0%	0.5%-1.0%	Benchmark

Note: The above data is based on GMO's internal models and public market data (2010-2020). Actual performance may fluctuate due to changes in market conditions.