One Screen, 13 Exchanges: Where Owner Earnings Yield Works Around the World

We ran the same Owner Earnings Yield screen on 13 global exchanges over 25 years. Same filters everywhere: OE Yield 5-50%, ROE > 10%, operating margin > 10%, market cap above the local threshold. Top 30 stocks, equal weight, annual rebalance in July.

Contents

1. Method
2. Full Results: All 13 Exchanges
3. The Winners: Nine Markets That Beat Their Local Benchmark
4. India: 17.17% CAGR, the clear leader
5. US: 12.25% CAGR, best risk-adjusted performance
6. Hong Kong: 10.44% CAGR, the biggest local-benchmark winner
7. China: 9.61% CAGR, +7.18% vs SSE Composite
8. South Africa (JSE): 9.64% CAGR, shallowest drawdown among outperformers
9. Sweden (STO): 9.59% CAGR, +7.03% vs OMX Stockholm 30
10. Taiwan: 6.20% CAGR, +2.11% vs TAIEX
11. Switzerland (SIX): 5.86% CAGR, +4.12% vs SMI
12. Japan (JPX): 4.86% CAGR, +1.54% vs Nikkei 225
13. The Underperformers: Four Markets Below Their Local Benchmark
14. Israel (TLV): 6.47% CAGR, -1.39% vs SPY
15. Saudi Arabia (SAU): 4.19% CAGR, -3.66% vs SPY
16. Korea (KSC): 3.73% CAGR, -1.62% vs KOSPI
17. Thailand (SET): 3.83% CAGR, -1.30% vs SET Index
18. Excluded Exchanges
19. Key Patterns Across Markets
20. Measuring against the right benchmark changes the picture
21. Universe size doesn't predict performance
22. Emerging markets delivered more absolute return, developed markets delivered more consistency
23. Drawdowns vary wildly
24. Cash periods signal trouble
25. Currency Note
26. What This Means for Investors
27. Part of a Series
28. Run This Screen Yourself
29. References

Six of the 13 exchanges beat the S&P 500 benchmark. But measured against local indices, 9 of 13 outperformed. The difference matters. Comparing a Swedish portfolio to SPY mixes currency effects with stock selection. Comparing it to the OMX Stockholm 30 isolates the strategy's contribution. India leads at 17.17% CAGR with +5.11% excess over the Sensex. The US delivers the best risk-adjusted result (0.432 Sharpe, zero cash periods). Hong Kong, Sweden, and China all produced over +7% annual alpha against their local benchmarks, numbers that were invisible when SPY was the only yardstick.

Method

Data source: Ceta Research (FMP financial data warehouse) Signal: Owner Earnings Yield = (Net Income + D&A - min(|Capex|, D&A)) / Market Cap Filters: OE Yield 5-50%, ROE > 10%, Operating Margin > 10% Portfolio: Top 30 by OE Yield, equal weight Rebalancing: Annual (July, after FY filings + 45-day lag) Execution: Next-day close (market-on-close after signal date) Cash rule: Hold cash if fewer than 10 stocks qualify Transaction costs: Size-tiered model (0.1-0.5% one-way based on market cap) Benchmarks: Local currency index where available (Sensex, S&P 500, Hang Seng, etc.), with SPY as fallback for exchanges without a reliable local index Period: 2000-2025 (25 annual periods)

Market cap thresholds vary by exchange. The US uses $1B USD. Smaller markets use lower thresholds calibrated to local liquidity. Full methodology: backtests/METHODOLOGY.md.

Full Results: All 13 Exchanges

SPY benchmark: 7.85% CAGR. Local benchmarks listed per exchange. JSE, TLV, and SAU fall back to SPY (no reliable local index in the dataset).

The Winners: Nine Markets That Beat Their Local Benchmark

India: 17.17% CAGR, the clear leader

India produced the highest absolute return of any exchange tested. +5.11% annual excess over the Sensex, with a max drawdown (-33.75%) shallower than most exchanges in the set. The Indian universe is deep. 28.0 stocks per rebalance means the screen almost always found enough qualifying names. Only 4 of 25 periods went to cash, all in the early years when FMP coverage was thinner.

The old SPY comparison showed +9.32% excess, but that mixed INR/USD currency effects with stock selection. Against the Sensex, the alpha is +5.11%. Still the strongest absolute performer, but the honest excess is roughly half what the SPY comparison suggested. The rupee depreciated roughly 2.7% annually against the dollar over this period, which inflated the SPY gap.

What makes India work? A large, growing pool of profitable mid-cap companies. Indian industrials, IT services, and consumer franchises frequently hit the ROE and margin filters while trading at high OE yields. The economic growth cycle concentrates value in the right sectors for this screen.

Full analysis: OE Yield on Indian Stocks (BSE + NSE)

US: 12.25% CAGR, best risk-adjusted performance

The US is the consistency story. 12.25% CAGR, 0.432 Sharpe (highest of all 13 exchanges), zero cash periods. The portfolio always found at least 10 qualifying stocks. The +4.40% excess over the S&P 500 is the same whether you measure against SPY or the local benchmark, because they're the same index.

OE Yield's golden era on US stocks was the early 2000s. While the dot-com bust wiped out growth stocks, the screen's filters naturally avoided overvalued names. From 2000 to 2005, OE Yield compounded at over 20% annually while SPY went sideways. The strategy's weakness is clear too: 2017-2023 underperformed as mega-cap growth dominated.

Full analysis: OE Yield on US Stocks (NYSE+NASDAQ+AMEX)

Hong Kong: 10.44% CAGR, the biggest local-benchmark winner

Hong Kong is the exchange where the local benchmark changes the picture most dramatically. The Hang Seng returned just 1.64% CAGR over this period. OE Yield's +8.80% excess is the largest of any exchange in the set. The old SPY comparison showed only +2.59%, understating the real alpha by more than 6 percentage points.

The HKD is pegged to the USD, so there's no currency translation effect. This +8.80% excess is real in both local and dollar terms.

Zero cash periods and 26.6 average stocks confirm a deep value universe. Hong Kong's property conglomerates, industrial holding companies, and Chinese H-share listings provide a steady supply of high-OE-yield names. The drawdown is steep (-46.75%), reflecting Hong Kong's sensitivity to mainland Chinese economic cycles and geopolitical risk.

Full analysis: OE Yield on Hong Kong Stocks (HKSE)

China: 9.61% CAGR, +7.18% vs SSE Composite

China is the third big winner of the local benchmark switch. The SSE Composite returned just 2.43% CAGR over this period, putting OE Yield's excess at +7.18%. The old SPY comparison showed only +1.76%. The 50.43% max drawdown is the second worst of any exchange tested (behind Japan), but the alpha against local equities is substantial.

China's A-share market is retail-dominated. Value signals work differently here. The OE Yield screen catches state-owned enterprises trading at low valuations, and while the volatility of Chinese equities is punishing, the strategy still produced meaningful alpha over the local market in most years.

Full analysis: OE Yield on Chinese Stocks (SHH+SHZ)

South Africa (JSE): 9.64% CAGR, shallowest drawdown among outperformers

JSE produced +1.78% excess over SPY (no local index available as benchmark) with the shallowest max drawdown of any outperforming exchange (-29.80%). The screen works well on South Africa's consumer, healthcare, and financial franchises. Mining companies mostly fail the margin and ROE filters, leaving a portfolio concentrated in high-quality domestic businesses.

The low Sharpe (0.035) reflects high local currency volatility and the 9% South African risk-free rate used in the calculation. The 5 cash periods also drag down risk-adjusted metrics. But on a pure return basis, JSE quietly outperforms most developed markets.

Sweden (STO): 9.59% CAGR, +7.03% vs OMX Stockholm 30

Sweden is the second big winner of the local benchmark switch. The OMX Stockholm 30 returned just 2.55% CAGR, giving OE Yield +7.03% excess. The old SPY comparison showed only +1.74%. A 0.390 Sharpe (second highest after the US), max drawdown of -40.56%, and an 80% win rate against the local index make Sweden one of the strongest results in the set.

Sweden's exchange is loaded with high-quality industrials (Atlas Copco, Sandvik, Alfa Laval) and tech companies (Hexagon, Ericsson) that frequently pass the quality filters. Only 4 cash periods, and the 26.0 average stocks per rebalance show consistent depth. For European investors looking for OE Yield exposure, Sweden is the strongest single-exchange option.

Full analysis: OE Yield on Swedish Stocks (STO)

Taiwan: 6.20% CAGR, +2.11% vs TAIEX

Taiwan was an "underperformer" in the old SPY-only comparison (-1.65% excess). Against the TAIEX, it's a clear winner at +2.11%. The strategy works here. It was the foreign benchmark that made it look bad.

Taiwan still has the best drawdown profile of any exchange tested: -20.78%. The screen selects Taiwanese electronics manufacturers and semiconductor supply chain companies with high margins and moderate valuations. 5 cash periods and a universe that rarely fills 30 positions limit absolute performance, but the 0.289 Sharpe is respectable.

Switzerland (SIX): 5.86% CAGR, +4.12% vs SMI

Switzerland flips from underperformer (-1.99% vs SPY) to strong outperformer (+4.12% vs SMI). The SMI returned just 1.74% CAGR over this period. Switzerland has the smallest average portfolio (17.0 stocks) of any exchange. Nestle, Roche, Novartis, and the banking giants rarely pass the OE Yield > 5% filter because they're too expensive. What remains is a thin set of Swiss industrials and specialty companies. Zero cash periods and +4.12% alpha show the screen picks well from what's available.

Japan (JPX): 4.86% CAGR, +1.54% vs Nikkei 225

Japan flips from the biggest disappointment (-2.99% vs SPY) to a modest outperformer (+1.54% vs the Nikkei 225). The Nikkei returned 3.31% CAGR, so the bar was lower than SPY. The alpha is thin but positive.

The universe is large (28.8 avg stocks, the highest of any exchange), so depth isn't the problem. The -55.50% max drawdown is the worst of all 13 exchanges. Japan's value trap problem is well documented: cross-shareholdings, poor capital allocation, entrenched management. But measured against the local market, the OE Yield screen at least avoided the worst of it.

The Underperformers: Four Markets Below Their Local Benchmark

Israel (TLV): 6.47% CAGR, -1.39% vs SPY

Israel underperforms SPY by a small margin (no local index available as benchmark). The 7 cash periods (28% of the backtest) are the primary drag. Israel's exchange is smaller, and many of its best companies dual-list on NASDAQ, pulling them out of the TLV universe. The -35.52% max drawdown is moderate. When the screen finds stocks, they hold up well. There just aren't always enough of them.

Saudi Arabia (SAU): 4.19% CAGR, -3.66% vs SPY

Saudi Arabia's stock exchange only opened to foreign investors in 2015. The earlier years of the backtest have limited coverage. 7 cash periods (28%) and a 0.037 Sharpe reflect a market where the OE Yield signal doesn't yet have enough depth or history to generate consistent alpha. No local index available, so SPY is the benchmark.

Korea (KSC): 3.73% CAGR, -1.62% vs KOSPI

Korea's market is dominated by Samsung, Hyundai, LG, and SK Group subsidiaries. Many of these are capital-intensive with complex cross-holdings that compress ROE below the 10% threshold. What passes the screen are smaller companies that lack the liquidity and institutional interest to drive price discovery. 7 cash periods and a 0.041 Sharpe tell the story. The KOSPI returned 5.35% CAGR, so even the local benchmark beats the screen.

Thailand (SET): 3.83% CAGR, -1.30% vs SET Index

Thailand is the weakest performer vs its local benchmark. The SET Index returned 5.13% CAGR. The Thai market has a large number of listed companies, but many are family-controlled businesses with low transparency. The OE Yield screen selects what passes the quality filters, but the valuation signal doesn't translate into forward returns in this market.

Excluded Exchanges

We tested seven additional exchanges but excluded them from the final comparison due to data quality issues:

The split artifact problem deserves explanation. FMP's adjusted close prices sometimes fail to account for stock splits correctly. When a stock splits 10:1, the historical price should be divided by 10. If it isn't, a $5 stock suddenly looks like it returned 900% in one year. One or two bad prices in a 30-stock portfolio can inflate annual returns by 20-50 percentage points and make the entire backtest unreliable. We drop any exchange where we can identify these artifacts rather than publish misleading numbers.

Key Patterns Across Markets

Measuring against the right benchmark changes the picture

Seven exchanges that underperformed SPY were actually compared against a foreign-currency benchmark. When measured against local indices, three of them flip from underperformer to outperformer: Taiwan (+2.11% vs TAIEX), Switzerland (+4.12% vs SMI), and Japan (+1.54% vs Nikkei). The SPY comparison mixes currency depreciation with stock selection. Local benchmarks isolate the strategy's actual contribution.

Universe size doesn't predict performance

Japan has the deepest universe (28.8 avg stocks) and the worst drawdown. Switzerland has the thinnest (17.0) and still produces +4.12% alpha vs its local index. India (28.0) and the US (23.2) both outperform despite different portfolio sizes. The signal's strength matters more than the number of stocks it finds.

Emerging markets delivered more absolute return, developed markets delivered more consistency

India (+5.11% vs Sensex), Hong Kong (+8.80% vs Hang Seng), and China (+7.18% vs SSE Composite) produced the highest excess returns against local benchmarks. But the US (0.432 Sharpe) and Sweden (0.390 Sharpe) delivered the best risk-adjusted performance. There's a tradeoff between absolute alpha and smoothness. Emerging market alpha comes with wider drawdowns and more volatile year-to-year returns.

Drawdowns vary wildly

The spread between the shallowest and deepest max drawdown is 35 percentage points. Taiwan (-20.78%) vs Japan (-55.50%). This isn't a function of the strategy. It's a function of the market. OE Yield inherits the tail risk of whatever exchange it runs on. Investors should size positions based on the max drawdown of the specific market, not the global average.

Cash periods signal trouble

Exchanges with 7 cash periods (Israel, Saudi Arabia, Korea) all underperform their benchmarks. When the screen can't find 10 qualifying stocks in more than a quarter of rebalance dates, the universe is too thin or the market structure doesn't produce enough high-quality value names. Cash drag alone doesn't explain the underperformance, but it correlates with it.

Currency Note

The "Excess vs Local" column in the table is currency-neutral. Both the portfolio and the benchmark are denominated in the same local currency, so exchange rate movements cancel out. This is the honest measure of whether the screen adds value on a given exchange.

The "SPY Excess" column still carries currency risk. A portfolio denominated in SEK, JPY, or INR is being compared to a USD benchmark. Currency depreciation inflates the gap for outperformers and deepens it for underperformers.

For three exchanges (JSE, TLV, SAU), no reliable local index was available. These fall back to SPY, meaning currency effects remain embedded in their results. South Africa is the most affected: the ZAR depreciated roughly 5% annually vs USD, so JSE's +1.78% excess over SPY may overstate the dollar-adjusted alpha.

What This Means for Investors

OE Yield is a global signal. It works on exchanges with very different market structures, from India's growth-oriented mid-cap universe to Sweden's quality-industrial complex to Hong Kong's property-and-holdco market. The mechanism is consistent: screen for businesses that generate high owner earnings relative to their price, confirm quality with ROE and margin filters, and hold.

Nine of 13 exchanges beat their local benchmark. That's a stronger result than the old "6 of 13 beat SPY" framing suggested. Taiwan, Switzerland, and Japan all produce positive alpha against their own markets. They only looked bad because they were being measured against a foreign index in a stronger currency.

Four markets still underperform: Korea, Thailand, Israel, and Saudi Arabia. Korea's chaebol concentration, Thailand's low transparency, and thin universes in Israel and Saudi Arabia all limit the screen's effectiveness.

The practical takeaway: the strongest results are Hong Kong (+8.80% vs Hang Seng), China (+7.18% vs SSE Composite), Sweden (+7.03% vs OMX30), India (+5.11% vs Sensex), and the US (+4.40% vs S&P 500). These five exchanges combine meaningful alpha with reasonable portfolio depth. Diversifying across all 13 exchanges without discrimination would dilute the strong results with weak ones.

Part of a Series

This comparison covers the full set of exchanges we tested. Dedicated deep-dives with year-by-year returns, screen SQL, and detailed analysis:

• OE Yield on US Stocks: 12.25% CAGR, 0.432 Sharpe, zero cash periods
• OE Yield on Indian Stocks (BSE + NSE): 17.17% CAGR, +5.11% vs Sensex
• OE Yield on Hong Kong Stocks (HKSE): 10.44% CAGR, +8.80% vs Hang Seng
• OE Yield on Swedish Stocks (STO): 9.59% CAGR, +7.03% vs OMX30
• OE Yield on Chinese Stocks (SHH+SHZ): 9.61% CAGR, +7.18% vs SSE Composite

Run This Screen Yourself

The backtest code is open source. Clone the repo, configure your API key, and run any exchange:

git clone https://github.com/ceta-research/backtests.git
cd backtests
pip install -r requirements.txt
python owner_earnings.py --exchange US --start 2000 --end 2025

Or run the live screen on any exchange via the web UI: cetaresearch.com/data-explorer

Get your API key at cetaresearch.com.

References

• Buffett, W. (1986). "Berkshire Hathaway Annual Letter to Shareholders."
• Greenwald, B., Kahn, J., Sonkin, P. & van Biema, M. (2001). Value Investing: From Graham to Buffett and Beyond. John Wiley & Sons.
• Greenblatt, J. (2006). The Little Book That Beats the Market. John Wiley & Sons.
• Fama, E. & French, K. (1998). Value versus Growth: The International Evidence. Journal of Finance.

Data: Ceta Research, FMP financial data warehouse. 13 exchanges, 25-year backtest (2000-2025). Annual rebalance (July), equal weight top 30, MOC execution. Transaction costs included. Local benchmarks where available, SPY (7.85% CAGR) as fallback.