Graham Number Timing in Sweden: Best Risk-Adjusted Returns Globally (0.544 Sharpe)

Benjamin Graham's intrinsic value formula delivered a 0.544 Sharpe ratio on Swedish stocks over 25 years. That's the highest risk-adjusted return across all 14 global exchanges we tested.

Contents

1. The Formula
2. Backtest Results (2000-2025, With Costs)
3. Why Sweden Excels at Risk-Adjusted Returns
4. When Graham Timing Works in Sweden
5. When Graham Timing Struggles in Sweden
6. The Quality Filters
7. Portfolio Construction
8. The Screen
9. The Academic Foundation
10. Why This Isn't a Free Lunch
11. Current Signal (March 2026)
12. Key Takeaways
13. Methodology

For context, a Sharpe ratio above 0.5 is exceptional for equity strategies. Sweden achieved 0.544 vs SPY's 0.354, that's 54% better risk-adjusted performance.

What makes Sweden special: the strategy delivered +4.97% annual alpha while capturing only 72% of downside risk. You get 113.7% of upside in bull markets and dodge 28% of losses in bear markets. That's the asymmetry every investor wants.

On $10,000 invested in 2000, you'd have $181,593 from Graham timing vs $62,745 from SPY, nearly 3x SPY's terminal value.

The Formula

Graham's fair value equation:

Graham Number = sqrt(22.5 × EPS × Book Value Per Share)

Where: - EPS = Net Income / Shares Outstanding - BVPS = Total Stockholders' Equity / Shares Outstanding - 22.5 = Graham's constant (P/E of 15 × P/B of 1.5)

The signal: Buy stocks where Price/Graham < 1.0

When Volvo trades at SEK 200 but its Graham Number is SEK 260, the Price/Graham ratio is 0.77, a 23% discount to intrinsic value.

Backtest Results (2000-2025, With Costs)

The strategy outperformed SPY by 4.97% annually. That's $118,848 more on a $10,000 investment over 25 years.

The Sharpe ratio of 0.544 is the headline metric. For every unit of volatility risk, the strategy delivered 0.544 units of excess return above the risk-free rate. SPY's 0.354 means the strategy's risk-adjusted performance is 54% better.

The down-capture ratio of 72.3% means when SPY had a losing quarter, this strategy lost only 72% as much on average. The up-capture of 113.7% means when SPY gained, the strategy captured an extra 13.7% upside.

This is defensive value investing. You get meaningful alpha with controlled downside exposure.

The 12.6% cash periods (13 of 103 quarters) reflect Nordic market discipline, when valuations stretch, the signal stops firing. Cash drag is the price of not chasing expensive stocks.

Why Sweden Excels at Risk-Adjusted Returns

Quality-biased corporate culture

Swedish companies emphasize profitability, dividends, and conservative balance sheets. The ROE > 10% filter naturally selects well-managed firms. Companies like Volvo, Ericsson, and Atlas Copco have multi-decade track records of consistent earnings and book value growth.

Defensive sector tilt

Many qualifying stocks come from Industrials, Consumer Staples, and Utilities, sectors with predictable cash flows. These sectors weathered 2008, 2011, and 2020 crashes better than cyclicals, explaining the 72% down-capture.

Lower volatility than US

20.18% annualized volatility (vs SPY's 16.97%) is modest for a single-country strategy. Swedish equities are less volatile than broader European indices due to stable governance and strong labor protections.

SEK currency stability

All returns are SEK-denominated. The Swedish krona is relatively stable vs USD compared to emerging market currencies, reducing volatility for foreign investors.

When Graham Timing Works in Sweden

The 2002-2003 recovery

After the dot-com crash, cheap Swedish industrials with intact earnings rebounded hard:

The strategy dodged most of the 2002 crash (-1.77% vs -19.92%) and captured the recovery. Combined two-year excess: +37.27%.

The 2009-2010 financial crisis recovery

Swedish stocks below Graham Number were positioned perfectly for the post-crisis rebound:

When global markets panicked in 2008-2009, defensive Swedish stocks with strong balance sheets outperformed once fear subsided.

2023-2024: Recent consistency

The most recent two years show the signal still working:

The 2024 outperformance (+13.43%) reflects value rotation after tech growth slowdown.

When Graham Timing Struggles in Sweden

Cash periods during expensive markets

The strategy went to cash 13 times (12.6% of quarters), primarily during valuation peaks:

• 2007-2008: Pre-crisis bubble, everything expensive
• 2015-2016: Negative interest rates inflated valuations
• 2021: Peak growth premium, few value stocks qualified

When cash, the strategy holds SPY (matches benchmark). These periods explain why total excess (+4.97%) is spread across 87% of the time actually invested.

Momentum markets

When growth stocks lead and value lags, the strategy underperforms:

The 2017-2019-2021 underperformance (-45.79% cumulative) required conviction to hold through. Value investing means watching growth double while your portfolio grinds sideways.

Max drawdown

-55.40% max drawdown (vs SPY's -45.53%) reflects value stocks getting cheaper during crashes. The 2008-2009 crisis drove Swedish industrials down alongside global markets.

The Quality Filters

We don't buy every Swedish stock below Graham Number. We add filters to avoid value traps:

1. ROE > 10%. Ensures profitability and capital efficiency
2. Net Income > 0. No loss-making companies
3. Total Equity > 0. No negative book value
4. Market Cap > SEK 5 billion (~$500M USD). Liquidity threshold

These filters eliminate distressed companies where low Price/Graham reflects permanent impairment (declining industries, zombie companies) rather than temporary mispricing.

Portfolio Construction

• Selection: Top 30 stocks by deepest discount to Graham Number (lowest Price/Graham ratio first)
• Weighting: Equal weight (3.33% per stock when 30 stocks qualify)
• Rebalancing: Quarterly (January, April, July, October)
• Transaction costs: 5-10 bps depending on market cap (round-trip, Nordic model)
• Exchange: Stockholm Stock Exchange (STO)

Quarterly rebalancing aligns with when new earnings data becomes available. Graham Number changes with reported net income and book value.

The 24.3 average holdings confirm the strategy runs below the 30-stock target most quarters. Some periods have 15-20 qualifying stocks, still well-diversified, but concentrated in high-conviction value names.

The 13 cash periods occur when fewer than 10 stocks qualify (typically in overheated markets). This is a feature, not a bug, it prevents buying expensive stocks just to stay invested.

The Screen

Run this on live Swedish stock data to see which stocks currently trade below their Graham Number:

WITH latest_filings AS (
    SELECT i.symbol,
           i.netIncome,
           b.totalStockholdersEquity,
           k.marketCap,
           k.returnOnEquity,
           ROW_NUMBER() OVER (PARTITION BY i.symbol ORDER BY i.dateEpoch DESC) AS rn
    FROM income_statement i
    JOIN balance_sheet b ON i.symbol = b.symbol
        AND ABS(CAST(i.dateEpoch AS BIGINT) - CAST(b.dateEpoch AS BIGINT)) < 86400 * 60
    JOIN key_metrics k ON i.symbol = k.symbol
        AND ABS(CAST(i.dateEpoch AS BIGINT) - CAST(k.dateEpoch AS BIGINT)) < 86400 * 60
    JOIN profile p ON i.symbol = p.symbol
    WHERE i.period IN ('FY', 'Q4')
      AND b.period IN ('FY', 'Q4')
      AND k.period IN ('FY', 'Q4')
      AND p.exchange = 'STO'
      AND i.netIncome > 0
      AND b.totalStockholdersEquity > 0
      AND k.returnOnEquity > 0.10
      AND k.marketCap > 5e9  -- SEK 5 billion (~$500M USD)
),
current_prices AS (
    SELECT symbol, adjClose AS price,
           ROW_NUMBER() OVER (PARTITION BY symbol ORDER BY dateEpoch DESC) AS rn
    FROM stock_eod
    WHERE dateEpoch > CAST(EXTRACT(EPOCH FROM CURRENT_DATE - INTERVAL '7 days') AS INTEGER)
),
graham_calc AS (
    SELECT
        f.symbol,
        f.netIncome,
        f.totalStockholdersEquity,
        f.marketCap,
        f.returnOnEquity,
        p.price,
        -- Graham Number = sqrt(22.5 * EPS * BVPS)
        SQRT(22.5 * (f.netIncome / (f.marketCap / p.price)) * (f.totalStockholdersEquity / (f.marketCap / p.price))) AS graham_number
    FROM latest_filings f
    JOIN current_prices p ON f.symbol = p.symbol AND p.rn = 1
    WHERE f.rn = 1 AND p.price > 0
)
SELECT
    symbol,
    ROUND(price, 2) AS current_price_sek,
    ROUND(graham_number, 2) AS graham_number,
    ROUND(price / graham_number, 2) AS price_to_graham_ratio,
    ROUND(returnOnEquity * 100, 1) AS roe_pct,
    ROUND(marketCap / 1e9, 2) AS market_cap_bn_sek
FROM graham_calc
WHERE price / graham_number < 1.0
ORDER BY price / graham_number ASC
LIMIT 30;

Run this screen on live data →

Stocks with price_to_graham_ratio < 1.0 are trading below Graham's fair value estimate. Lower ratios = deeper discounts.

The Academic Foundation

Benjamin Graham (1949): The Intelligent Investor introduced the Graham Number as a conservative valuation metric for defensive investors. Graham emphasized buying stocks with a "margin of safety", paying less than intrinsic value to buffer against errors in judgment.

Asness, Moskowitz, Pedersen (2013): "Value and Momentum Everywhere" showed value strategies work across asset classes and geographies. Value premiums are persistent globally, though they vary by market structure and investor behavior.

Fama & French (1998): "Value Versus Growth: The International Evidence" showed value stocks outperformed growth stocks in 12 of 13 international markets, earning 7.68% annual premiums on average. Sweden's +4.97% excess aligns with this global pattern.

Sweden's 0.544 Sharpe ratio reflects what happens when value investing meets a market with strong corporate governance, defensive sector mix, and disciplined valuation cycles.

Why This Isn't a Free Lunch

You pay with cash drag

12.6% cash periods mean missing some market returns when Graham signals don't fire. The strategy adapts to market conditions, sometimes that means sitting out.

You pay with underperformance in momentum markets

The 2017-2021 period saw -45.79% cumulative underperformance across three years. Value investing means watching growth stocks double while your portfolio grinds sideways.

You pay with drawdowns

-55.40% max drawdown means watching SEK 100k turn into SEK 45k. That's the price of admission for value strategies during crashes.

You need patience

Value investing means enduring 3-5 year periods where the market ignores fundamentals. The strategy works over decades, not quarters.

Currency risk for foreign investors

All returns are SEK-denominated. Dollar-based investors face additional USD/SEK exchange rate risk (krona fluctuates ±10-15% vs USD in volatile years).

Current Signal (March 2026)

As of the latest data, check the live screen above for the current count of Swedish stocks trading below Graham Number with ROE > 10%.

The signal adapts to market conditions. In expensive markets, the count drops below 10 (go to cash). In crashes, the count spikes above 30 (buy aggressively).

Key Takeaways

What worked: - +4.97% annual alpha over 25 years (second best globally) - 0.544 Sharpe ratio (best risk-adjusted performance globally) - 72.3% down-capture (excellent downside protection) - 113.7% up-capture (amplified gains in bull markets) - 20.18% volatility (controlled risk for single-country strategy)

What didn't: - -55.40% max drawdown (value stocks crash too) - 12.6% cash periods (signals don't always fire) - -45.79% cumulative underperformance 2017-2021 (growth mania)

Who this is for: - Long-term investors (5+ year horizon) - Risk-conscious value investors (want alpha with controlled volatility) - Nordic market exposure (can stomach 20% volatility) - SEK-denominated returns acceptable (or willing to hedge currency)

Who this isn't for: - Short-term traders (mean reversion takes time) - Growth-focused investors (misses momentum rallies) - Maximum-return seekers (India's +5.43% beats Sweden's +4.97%)

Methodology

• Universe: Stockholm Stock Exchange (STO)
• Period: 2000-01-01 to 2025-10-01 (103 quarters, 25 years)
• Rebalancing: Quarterly (Jan 1, Apr 1, Jul 1, Oct 1)
• Position sizing: Equal weight, 30 stocks max, 10 minimum (cash if fewer qualify)
• Transaction costs: Nordic model (5-10 bps depending on market cap)
• Currency: SEK-denominated returns
• Survivorship bias: Controlled (includes delisted stocks)
• Benchmark: SPY total return (for global comparison)

Full backtest code: github.com/ceta-research/backtests/graham-timing

Part of a Series: Global | US | India | Global | US | Switzerland | Sweden | Japan

Run It Yourself

Explore the data behind this analysis on Ceta Research. Query our financial data warehouse with SQL, build custom screens, and run your own backtests across 70,000+ stocks on 20 exchanges.

Data: Ceta Research | FMP fundamental data warehouse | 25-year backtest (2000-2025)