Low Debt Quality Stocks: Canada (TSX) — 25-Year Backtest
Low-debt, high-quality Canadian stocks returned 9.16% CAGR over 25 years vs 7.83% for the S&P 500. Total return: 795.28% vs 558.69%. But the real story isn't the outperformance on the upside. It's what happened when markets fell.
Contents
- Method
- The Signal
- The Screen (SQL)
- What We Found
- Annual Returns (July to July)
- Bear Markets: Where Canada Shines
- The Tech Bubble Hypothesis
- The Lost Decade (2011–2022)
- 2024: The Recovery
- Backtest Methodology
- Limitations
- Takeaway
- Part of a Series
- Run This Screen Yourself
Down capture: -13.01%. Negative. When SPY dropped, this portfolio tended to rise. That's the signal worth understanding.
Method
- Universe: TSX-listed stocks with market cap above CAD 700M, actively trading.
- Period: July 2000 to July 2025 (25 annual holding periods).
- Rebalance: Annual, each July. Equal weight across all qualifying stocks.
- Benchmark: S&P 500 Total Return (SPY), in USD. Portfolio returns in CAD.
- Cash rule: Hold cash if fewer than 10 stocks qualify. Canada had 0 cash periods, averaging 35 stocks per holding period.
- Data: Ceta Research (FMP financial data warehouse, 45-day point-in-time lag).
The Signal
Two filters combine to define the portfolio.
D/E < 0.5 (Debt/Equity from annual financial statements). This isn't just a conservative cut. It's a structural filter. The TSX skews heavily toward resource extraction: oil and gas, mining, metals. These industries need capital. They borrow. A D/E threshold of 0.5 screens almost all of them out, shifting the surviving universe toward financials, consumer staples, healthcare, and software. That's the Canada story.
Piotroski F-Score >= 7 (out of 9 possible points). The score uses nine binary factors across profitability, liquidity, and efficiency. A stock must pass at least seven to qualify:
- F1: Net income > 0
- F2: Operating cash flow > 0
- F3: Return on assets improved year-over-year
- F4: Operating cash flow > net income (cash quality check)
- F5: Long-term debt ratio decreased
- F6: Current ratio improved
- F7: Shares outstanding didn't increase (no dilution)
- F8: Asset turnover improved
- F9: Gross margin improved
A score of 7+ means the business is profitable, liquid, not diluting shareholders, and improving across most operating dimensions. Combined with D/E < 0.5, the filter is strict. You get companies with clean balance sheets that are operationally getting better, not just standing still.
The Screen (SQL)
Run this on current data to see today's qualifying Canadian stocks:
SELECT p.exchange, r.symbol, p.companyName, p.sector,
ROUND(r.debtToEquityRatioTTM, 3) AS de_ratio,
ROUND(k.returnOnEquityTTM * 100, 1) AS roe_pct,
ROUND(r.operatingProfitMarginTTM * 100, 1) AS opm_pct,
ROUND(r.interestCoverageRatioTTM, 1) AS interest_coverage,
ROUND(k.freeCashFlowYieldTTM * 100, 2) AS fcf_yield_pct,
ROUND(k.marketCap / 1e9, 2) AS mktcap_b
FROM financial_ratios_ttm r
JOIN key_metrics_ttm k ON r.symbol = k.symbol
JOIN profile p ON r.symbol = p.symbol
WHERE r.debtToEquityRatioTTM >= 0
AND r.debtToEquityRatioTTM < 0.50
AND k.returnOnEquityTTM > 0.08
AND r.operatingProfitMarginTTM > 0.08
AND r.interestCoverageRatioTTM > 5.0
AND p.isActivelyTrading = true
AND k.marketCap > 700000000
AND p.exchange IN ('TSX')
ORDER BY de_ratio ASC
LIMIT 30
What We Found
| Metric | Portfolio | SPY |
|---|---|---|
| CAGR | 9.16% | 7.83% |
| Total Return | 795.28% | 558.69% |
| Max Drawdown | -31.65% | -36.27% |
| Sharpe Ratio | 0.356 | — |
| Volatility | 18.7% | — |
| Down Capture | -13.01% | — |
| Up Capture | 82.35% | — |
| Win Rate | 48% | — |
| Cash Periods | 0/25 | — |
| Avg Stocks | 35 | — |
Annual Returns (July to July)
| Year | Portfolio (CAD) | SPY (USD) | Excess |
|---|---|---|---|
| 2000 | +32.2% | -14.8% | +47.0% |
| 2001 | +10.8% | -20.8% | +31.6% |
| 2002 | -5.0% | +3.3% | -8.3% |
| 2003 | +18.2% | +16.4% | +1.8% |
| 2004 | +37.3% | +7.9% | +29.4% |
| 2005 | +31.6% | +8.9% | +22.7% |
| 2006 | +21.1% | +20.9% | +0.2% |
| 2007 | +1.0% | -13.7% | +14.7% |
| 2008 | -14.0% | -26.1% | +12.1% |
| 2009 | +30.9% | +13.4% | +17.5% |
| 2010 | +33.5% | +32.9% | +0.6% |
| 2011 | -12.3% | +4.1% | -16.4% |
| 2012 | -0.7% | +20.9% | -21.5% |
| 2013 | +30.8% | +24.5% | +6.3% |
| 2014 | -5.2% | +7.4% | -12.6% |
| 2015 | -1.0% | +3.4% | -4.4% |
| 2016 | +17.6% | +17.7% | -0.1% |
| 2017 | +12.0% | +14.3% | -2.4% |
| 2018 | -19.6% | +10.9% | -30.4% |
| 2019 | -14.8% | +7.1% | -21.9% |
| 2020 | +23.0% | +40.7% | -17.7% |
| 2021 | -18.9% | -10.2% | -8.8% |
| 2022 | +7.7% | +18.3% | -10.6% |
| 2023 | +16.9% | +24.6% | -7.7% |
| 2024 | +35.1% | +14.7% | +20.4% |
Bear Markets: Where Canada Shines
The dot-com bust is the clearest example. From July 2000 to July 2001, SPY fell 14.8% then another 20.8%. Canada's low-debt portfolio returned +32.2% then +10.8%. The gap across those two years: +78.6 percentage points combined.
2008 was a compressed version of the same story. SPY fell 26.1%. Canada fell only 14.0%. That -12.1% swing matters over time, particularly when compounded.
The -13.01% down capture figure summarizes it cleanly. Across all years SPY was negative, Canada didn't just fall less. It actually gained on average. That's a genuine portfolio characteristic, not noise.
The Tech Bubble Hypothesis
The D/E filter's effect on the TSX composition explains much of this. Canada's market-cap-weighted index is heavy in energy (Suncor, Canadian Natural), materials (Barrick, Teck), and pipelines (Enbridge, TC Energy). All of these tend to carry meaningful leverage by industry necessity. Screening them out leaves the TSX's underrepresented sectors: Canadian banks (which do have D/E above 0.5 due to deposit leverage), insurance, consumer goods, and technology. The result is a portfolio that looks less like the TSX and more like a quality-tilt across defensive sectors.
The Lost Decade (2011–2022)
Eleven years of erratic performance followed the strong early run. 2011 through 2013 saw two down years and a near-flat year. 2018 and 2019 were particularly painful: -19.6% and -14.8% in consecutive periods, a combined -31.2% draw before recovery. The CAD/USD currency relationship added friction. During risk-off periods when the USD strengthened, CAD-denominated returns looked worse in USD terms even when underlying businesses held up.
The patient investor who stayed the course through 2024 was rewarded: +35.1%, the best single year in the backtest.
2024: The Recovery
2024's +35.1% was 20.4 percentage points ahead of SPY. This came after a rough 2021-2023 stretch where the portfolio lagged on five of six metrics. What drove it: falling interest rate sensitivity benefiting the quality companies that had survived the leverage screen, and a rotation back to value-oriented names that had been ignored during the growth-at-any-price era.
Backtest Methodology
| Parameter | Value |
|---|---|
| Universe | TSX, market cap > CAD 700M |
| Period | July 2000 – July 2025 |
| Rebalance | Annual, July |
| Weighting | Equal weight |
| D/E threshold | < 0.5 (annual FY statements) |
| Piotroski minimum | F-Score >= 7 |
| Cash rule | < 10 qualifying stocks → hold cash |
| Benchmark | SPY (S&P 500 Total Return) |
| Returns | Portfolio in CAD, SPY in USD |
| Data lag | 45-day point-in-time |
| Data source | Ceta Research / FMP warehouse |
Limitations
Currency mismatch. Portfolio returns are in CAD. SPY returns are in USD. This is an apples-to-oranges comparison for absolute levels, but it's the right comparison if you're a Canadian investor whose expenses are in CAD. In years when CAD weakened against USD, the portfolio looked worse versus SPY than the underlying business performance would suggest. The opposite also applies.
No transaction costs. Switching 35 stocks once per year has friction. Bid-ask spreads, brokerage commissions, and market impact aren't modeled. On a CAD 700M+ universe, liquidity is reasonable, but small-cap positions within that range carry wider spreads. Real-world returns would be slightly lower.
Survivorship bias. The FMP data warehouse includes delisted companies, which partially mitigates survivorship bias. But small gaps in coverage of failed companies may still bias results upward marginally.
Benchmark currency. SPY is USD. The portfolio is CAD. A cleaner benchmark for a Canadian investor would be the S&P/TSX Composite Total Return. Against that index, the excess return picture could look different, since the TSX had strong periods (notably commodity cycles) where this portfolio's leverage filter would have left it behind.
Point-in-time lag. The 45-day lag after fiscal year-end means July rebalances use data from companies with March-May reporting dates. Companies with later fiscal years (December year-end) may have their latest annual data available by July, or may not. This creates some timing inconsistency across holdings.
Takeaway
Canada doesn't outperform because of aggressive growth bets. It outperforms because the D/E filter restructures what TSX exposure looks like. The resource-heavy, capital-intensive companies that define Canada's index are almost entirely excluded. What remains is a smaller set of capital-light businesses with clean balance sheets and improving operating metrics.
The -13.01% down capture is the number that matters most. A portfolio that rises when markets fall doesn't just protect wealth. It compounds from a higher base when the recovery comes. Over 25 years, that effect shows up as 795.28% total return vs 558.69% for SPY.
Part of a Series
This post is part of the Low Debt Quality backtest series, tested across 13 markets:
Run This Screen Yourself
The SQL above runs on live TTM data. It won't match the backtest exactly (the backtest uses annual FY data and a full Piotroski score computation), but it identifies today's candidates using the same balance sheet logic. Paste it into the Ceta Research data explorer when available.
Data: Ceta Research, FMP financial data warehouse. Returns in CAD. Annual rebalance July, equal weight, 2000–2025.
