Deleveraging Tested on 13 Markets: Positive Excess Across All Exchanges

Contents

1. How the Strategy Works
2. Full Results: 13 Exchanges (2001-2025)
3. Data Quality Note: The FY2012 Filter
4. The Top Performers: India and Canada
5. India: 15.82% CAGR, +4.70% Excess
6. Canada: 11.28% CAGR, +6.20% Excess
7. US: The Largest Improvement from Clean Data
8. United States: 11.22% CAGR, +3.20% Excess
9. The Middle Ground: Sweden, Germany, Japan, Switzerland
10. UK and Hong Kong: Positive Signal, Structural Challenges
11. UK: 6.63% CAGR, +5.28% Excess vs FTSE
12. Hong Kong: 5.87% CAGR, +4.10% Excess vs HSI
13. Asia-Pacific: Taiwan, China, Korea, Thailand
14. Check Today's Global Count
15. Why the Signal Works Globally
16. Rate Regime Considerations
17. Strategy Limitations
18. What This Tells You

We ran the same deleveraging screen on 13 global stock exchanges. The signal - companies cutting their debt-to-equity ratio by 10%+ year-over-year while maintaining ROE above 8% - produced positive excess returns versus local benchmarks in all 13 markets.

India returned 15.82% CAGR, beating Sensex by 4.70pp. Canada returned 11.28%, a 6.20% edge over TSX Composite. Sweden added 5.88% vs OMX Stockholm. Japan beat Nikkei by 4.38pp. The US posted 11.22% CAGR, a 3.20% excess over SPY. Even markets with modest absolute returns - Switzerland, Hong Kong, Korea - showed positive signals when measured against their local benchmarks.

This is a universally profitable factor when properly implemented with clean data. The variations across markets reflect local efficiency, sector composition, and rate regimes - not signal failure.

How the Strategy Works

The screen is straightforward. Each year, we identify companies where:

• D/E ratio declined 10%+ from the prior fiscal year
• Prior D/E > 0.1, current D/E > 0.01 (excludes erroneous zero-leverage data)
• ROE > 8% (profitability filter)
• Market cap above the exchange-appropriate threshold

We select the top 30 qualifying stocks by magnitude of D/E reduction, weight them equally, and rebalance annually each July. Annual filing data with a 45-day lag prevents look-ahead bias. Transaction costs scale by market cap tier: 0.1% for large caps, 0.3% for mid, 0.5% for small.

All returns are in local currency. SPY benchmark is 8.02% CAGR over the same period.

Full Results: 13 Exchanges (2001-2025)

Note: Excess measured vs local benchmark (not SPY). Local currency returns.

Data: Ceta Research (FMP financial data warehouse). Transaction costs applied. Local currency returns.

Data Quality Note: The FY2012 Filter

Earlier versions of this analysis showed several markets (US, UK, Hong Kong) with negative excess returns. The reason: FMP's FY2012 financial data contained systematic errors where 29.62% of stocks reported zero D/E ratios despite having non-zero debt and equity.

The filter de_current > 0.01 excludes these erroneous entries. This cleaned the selection universe and revealed the true signal: deleveraging with maintained profitability is universally profitable when measured against local benchmarks.

The US CAGR improved from 4.07% to 11.22% (+3.20% excess, previously -3.95%). All 13 exchanges now show positive excess returns. The narrative changes from "regime-dependent failure" to "universally profitable signal with market-specific variations."

The Top Performers: India and Canada

India: 15.82% CAGR, +4.70% Excess

India's excess return comes with a down capture of 73.5% vs Sensex - when the Sensex fell in a given quarter, the Indian portfolio fell about 74 cents on the dollar. That still beats the benchmark. Compared against SPY, the asymmetry is even stronger: the portfolio falls roughly 41 cents for every dollar SPY loses, reflecting that Indian equities and US equities often diverge in downturns.

Why does the signal work here? Indian capital markets are less informationally efficient than the US or UK. When an Indian company publishes results showing 10%+ D/E reduction alongside maintained profitability, that information takes time to price fully. The stock continues re-rating for quarters after the filing. In New York or London, the same information is absorbed within days.

Indian companies also carry higher baseline leverage than Western peers, and debt access is more constrained by banking structure and credit availability. A company that cuts its D/E ratio while maintaining ROE above 8% is demonstrating real operational discipline - not just riding a rate cycle. Markets there reward that discipline with a slower, more durable re-rating.

The 21% cash rate (roughly 5 years in cash over the 25-year period) reflects periods where the qualifying universe was too thin. The long-run 15.82% CAGR holds up despite those gaps. For a dedicated analysis of the India results, see our India Deleveraging backtest post.

Canada: 11.28% CAGR, +6.20% Excess

Canada's result is structurally strong and has the best risk-adjusted metrics of any market. The TSX is heavily weighted toward resource companies and financials - sectors where balance sheet health gets close scrutiny from both equity and debt investors. A resource company that reduces its D/E ratio during a commodity downturn is signaling operational resilience that the market prices in over time.

Canada's 0% cash rate means the strategy stayed fully invested for the entire period. 23.6 average stocks gives solid diversification. The Sharpe ratio of 0.537 is the best of any market in this analysis, including India. The strategy returned more per unit of volatility in Canada than anywhere else.

The down capture of 54.5% vs TSX shows capital protection working as intended - the portfolio absorbed about half of every benchmark drawdown. That's the whole premise of a deleveraging strategy, and Canada is the market where it delivers most consistently.

For a full breakdown of the Canada results, see our Canada Deleveraging backtest post.

US: The Largest Improvement from Clean Data

United States: 11.22% CAGR, +3.20% Excess

Earlier results showed the US at 4.07% CAGR with -3.95% excess vs SPY. The FY2012 data errors contaminated the selection universe, pulling in companies with erroneous zero D/E ratios that were actually distressed or data-incomplete.

With the de_current > 0.01 filter, the US result transforms: 11.22% CAGR, +3.20% annual excess over SPY, Sharpe 0.418. The down capture improves from 125.8% (amplifying losses) to 103.7% (roughly tracking SPY losses). This matches the global pattern: companies reducing leverage while maintaining profitability earn positive excess when properly selected.

The US signal works across rate regimes when the data is clean. The pre-2013 vs post-2013 split that earlier analysis showed likely reflected contaminated data periods rather than a true regime break. With clean data, the signal produces consistent positive alpha.

For detailed US results, see our US Deleveraging backtest post.

The Middle Ground: Sweden, Germany, Japan, Switzerland

These four markets show CAGRs between 4-9%, but all four have meaningful positive excess returns when measured against their local benchmarks.

Sweden (+5.88% vs OMX): 9.05% CAGR, Sharpe 0.349, 13% cash. The strongest excess return in this group. Sweden has well-developed corporate governance and transparent reporting - when a company genuinely reduces leverage, the market takes time to price it because analysts aren't watching as closely as in the US. That delay creates the alpha window.

Japan (+4.38% vs Nikkei): 7.78% CAGR, Sharpe 0.355, 7% cash. Japan is interesting because corporate deleveraging has been a structural theme since the early 2000s - companies actively cleaned up post-bubble balance sheets for decades. The Nikkei has a low long-run CAGR (3.40%), which amplifies the relative edge. Post-2013 monetary easing (Japan's QE predates and exceeds the US in scale) compressed but didn't eliminate the signal.

Germany (+3.57% vs DAX): 8.69% CAGR, Sharpe 0.304, 0% cash, 16.0 average stocks. Solid and consistent. German industrial companies operate in a culture where conservative financing is standard, so the signal selects from an already-disciplined pool. The edge is real and persistent.

Switzerland (+2.06% vs SMI): 4.17% CAGR, Sharpe 0.178, 0% cash, 13.9 average stocks. Positive excess despite low absolute CAGR. The thin universe (13.9 stocks) is the main risk - the SMI is dominated by a handful of mega-caps, so the strategy often runs in a limited pool. Still, the signal produces consistent outperformance.

UK and Hong Kong: Positive Signal, Structural Challenges

UK: 6.63% CAGR, +5.28% Excess vs FTSE

The UK shows strong excess vs its local FTSE benchmark (+5.28%), but lags SPY by about 1.4 percentage points annually. This reflects UK equities' long underperformance vs US equities over this period, not signal failure.

The Sharpe ratio of 0.159 is positive, and 13.0 average stocks is thin for a large developed market - idiosyncratic risk is higher and the results less statistically reliable. The UK is best treated as a solid positive signal in local terms.

Hong Kong: 5.87% CAGR, +4.10% Excess vs HSI

Hong Kong's transformation is dramatic. Earlier contaminated data showed -0.70% CAGR and -2.47% excess. Clean data reveals 5.87% CAGR and +4.10% excess vs Hang Seng.

The HKSE has unique structural characteristics: high concentration of state-linked Chinese enterprises, capital flow dynamics that respond to mainland policy, and liquidity events. With proper data filtering, the deleveraging signal still works - companies reducing debt while maintaining profitability outperform the local benchmark, even in this complex environment.

The Sharpe of 0.096 and down capture of 92.0% show this is a volatile, drawdown-heavy strategy. But the positive excess is real.

Asia-Pacific: Taiwan, China, Korea, Thailand

Taiwan (+1.95% vs TAIEX): 6.32% CAGR, Sharpe 0.284, 24% cash. Positive excess with significant cash periods reflecting a thinner qualifying universe. Taiwan's semiconductor supply chain firms have different capital structures than traditional deleveraging targets.

China (+1.48% vs SSE): 5.67% CAGR, Sharpe 0.099, 0% cash. China's result includes extraordinary years: +155% in 2007 (Shanghai bubble), +104% in 2009. Against the local SSE benchmark, China shows +1.48% excess. Volatile but profitable.

Korea (+0.08% vs KOSPI): 4.89% CAGR, essentially matching KOSPI. Korea's 28% cash rate reflects chaebols - the structurally leveraged conglomerates that dominate the KSC - consistently failing the D/E reduction filter. When the signal finds qualifying names, it roughly matches the market. Neutral, not negative.

Thailand (+1.13% vs SET): 4.89% CAGR, small positive excess. Thailand is effectively a modest positive against its local benchmark over 25 years. With 17% cash periods and 20.6 average stocks when invested, the signal finds qualifying names but generates only a thin edge.

Check Today's Global Count

How many stocks currently qualify across exchanges? Run the query:

WITH current_fy AS (
    SELECT symbol, debtToEquityRatio AS de_current
    FROM financial_ratios
    WHERE period = 'FY'
      AND debtToEquityRatio IS NOT NULL
      AND date >= CAST(CURRENT_DATE::DATE - INTERVAL '18 months' AS VARCHAR)
    QUALIFY ROW_NUMBER() OVER (PARTITION BY symbol ORDER BY date DESC) = 1
),
prior_fy AS (
    SELECT symbol, debtToEquityRatio AS de_prior
    FROM financial_ratios
    WHERE period = 'FY'
      AND debtToEquityRatio IS NOT NULL
      AND date >= CAST(CURRENT_DATE::DATE - INTERVAL '30 months' AS VARCHAR)
      AND date < CAST(CURRENT_DATE::DATE - INTERVAL '12 months' AS VARCHAR)
    QUALIFY ROW_NUMBER() OVER (PARTITION BY symbol ORDER BY date DESC) = 1
),
km AS (
    SELECT symbol, returnOnEquityTTM AS roe, marketCap
    FROM key_metrics_ttm
)
SELECT
    p.exchange,
    COUNT(*) AS qualifying_stocks
FROM current_fy c
JOIN prior_fy pr ON c.symbol = pr.symbol
JOIN km k ON c.symbol = k.symbol
JOIN profile p ON c.symbol = p.symbol
WHERE pr.de_prior > 0.1
  AND c.de_current > 0.01
  AND (c.de_current - pr.de_prior) / pr.de_prior < -0.10
  AND k.roe > 0.08
GROUP BY p.exchange
ORDER BY qualifying_stocks DESC

Run this query live: cetaresearch.com/data-explorer?q=9X6rtMVU-M

For the current US qualifying stock list: cetaresearch.com/data-explorer?q=jc1D3eHgRx

Why the Signal Works Globally

The split between stronger and weaker performers isn't random. It follows a consistent pattern:

Works best where leverage discipline is scarce or structurally meaningful. India: debt is harder to access, and companies that earn the right to reduce it are demonstrating genuine financial progress. Canada: resource sector balance sheets are closely watched by credit markets, and operational improvement is genuinely informative.

Solid performance in developed markets with clean data. US, Germany, Sweden, Japan, Switzerland, and the UK all show positive excess returns. The signal captures companies genuinely improving their financial positions, and markets price that improvement even in efficient markets.

Weaker but still positive in Asia-Pacific. Hong Kong, Taiwan, China, Korea, and Thailand all show positive excess returns, though margins are thinner. Political and capital flow dynamics, opaque financing structures, and conglomerate-dominated markets reduce but don't eliminate the signal's power.

Quality filter is critical. The de_current > 0.01 filter removes data artifacts. The ROE > 8% filter separates voluntary deleveraging (operational strength) from forced deleveraging (distress). Both are essential for the signal to work globally.

Rate Regime Considerations

Earlier contaminated data suggested a pre-2013 vs post-2013 regime break, with ZIRP (zero-interest-rate policy) making deleveraging unprofitable. Clean data shows no such break.

The signal works across rate environments when properly filtered. Companies reducing debt while maintaining profitability create value whether rates are 0% or 5%. What changed in contaminated data was selection bias (picking distressed companies with data errors), not regime shift.

That said, rate environments do affect magnitude. Higher rate environments may amplify the signal's value (debt is expensive, deleveraging more meaningful). Lower rate environments may compress it (debt is cheap, deleveraging less differentiated). But the direction is consistent: positive excess across all tested markets and time periods.

Strategy Limitations

Currency differences. All returns are local currency. India's 15.82% in rupees includes historical rupee depreciation of 2-4% per year against the dollar. Canada's 11.28% is in CAD. Cross-currency comparisons require adjusting for expected exchange rate trends.

Data availability varies by exchange. Earlier years (2001-2004) have thinner FMP coverage on several exchanges, particularly India. The effective backtest start date with robust data is closer to 2005-2006 for some markets. Returns in the early period may reflect a thinner qualifying universe.

Annual filing frequency. The strategy uses annual fiscal year data with a 45-day lag. For companies that report quarterly, the signal can be 12 months stale by the time the next annual filing arrives. A company can deleverage in one year and re-lever in the next before the screen captures the reversal.

Korea and Taiwan cash rates. Korea's 28% cash rate reflects chaebols - the structurally leveraged conglomerates that dominate the KSC - consistently failing the D/E reduction filter. Taiwan's 24% cash reflects a similar dynamic in its semiconductor supply chain firms. Both markets have enough invested periods to show positive signals, but expect periods with thin universes.

China context. China's 5.67% CAGR includes extraordinary years: +155% in 2007 (Shanghai bubble), +104% in 2009. Despite those outlier years, China shows a +1.48% excess vs SSE. This is comparison data - the signal's behavior in Chinese markets is volatile and regime-dependent in ways the simple annual backtest doesn't fully capture.

What This Tells You

Use the deleveraging signal selectively. It works everywhere when measured against local benchmarks, but the magnitude varies by market efficiency, sector composition, and capital structure norms.

If you're building a strategy for India or Canadian equities, D/E reduction combined with an ROE filter is a strong alpha source. If you're focused on US, UK, or European markets, the signal produces solid positive excess with lower volatility. Asia-Pacific markets show positive but thinner edges.

The data quality lesson is critical: filtering out erroneous zero-leverage entries (de_current > 0.01) transforms the signal from apparently broken to universally profitable. This is not optional - it's the difference between selection bias and signal extraction.

Data: FMP financial data warehouse, 2000–2025. Updated April 2026.

Data: Ceta Research (FMP financial data warehouse). Backtest period: 2001-2025. Transaction costs applied. Full methodology at cetaresearch.com.