EV/EBITDA Sector-Relative Value on Canada: An Honest Assessment

We ran a sector-relative EV/EBITDA screen on 24 years of Canadian stock data from the TSX. The signal: buy when a stock's EV/EBITDA falls below 70% of its sector median, filtered for quality. The result: 8.42% CAGR vs 7.64% SPY. On the surface, that's positive. But the max drawdown of -44.68% is worse than SPY's -34.90%, and the win rate is only 44%. The strategy squeezes out marginal outperformance while carrying more tail risk than the benchmark. Canada is a useful case study in when sector-relative signals work and when they don't.

Contents

1. Method
2. The Screen
3. Sector-Relative EV/EBITDA Screen for TSX (SQL)
4. What We Found
5. 24 invested years. 44% win rate. MaxDD worse than the index.
6. What happened in the early years
7. Year-by-year returns
8. 2008: the worst year
9. 2012-2015: four years of consistent underperformance
10. 2016: +24.9% excess, the commodity rebound
11. Why Canada's resource composition is a structural problem
12. Backtest Methodology
13. Limitations
14. Conclusion

Method

Financial data sourced from key_metrics_ttm for EV/EBITDA and market cap, financial_ratios_ttm for debt metrics, profile for sector classification. 45-day lag on all financial statements to prevent look-ahead bias.

The Screen

Sector-Relative EV/EBITDA Screen for TSX (SQL)

WITH universe AS (
 SELECT k.symbol, p.companyName, p.exchange, p.sector,
 k.evToEBITDATTM AS ev_ebitda, k.returnOnEquityTTM AS roe,
 fr.debtToEquityRatioTTM AS de, k.marketCap
 FROM key_metrics_ttm k
 JOIN financial_ratios_ttm fr ON k.symbol = fr.symbol
 JOIN profile p ON k.symbol = p.symbol
 WHERE k.evToEBITDATTM BETWEEN 0.5 AND 25
 AND k.returnOnEquityTTM > 0.08
 AND (fr.debtToEquityRatioTTM IS NULL OR (fr.debtToEquityRatioTTM >= 0 AND fr.debtToEquityRatioTTM < 2.0))
 AND k.marketCap > 500000000
 AND p.sector IS NOT NULL
 AND p.exchange IN ('TSX')
),
sector_medians AS (
 SELECT exchange, sector,
 PERCENTILE_CONT(0.5) WITHIN GROUP (ORDER BY ev_ebitda) AS median_ev_ebitda,
 COUNT(*) AS n_sector_stocks
 FROM universe GROUP BY exchange, sector HAVING COUNT(*) >= 5
)
SELECT u.symbol, u.companyName, u.exchange, u.sector,
 ROUND(u.ev_ebitda, 2) AS ev_ebitda_ttm,
 ROUND(sm.median_ev_ebitda, 2) AS sector_median_ev_ebitda,
 ROUND(u.ev_ebitda / sm.median_ev_ebitda, 3) AS ev_ratio_to_sector,
 ROUND((1 - u.ev_ebitda / sm.median_ev_ebitda) * 100, 1) AS discount_pct,
 ROUND(u.roe * 100, 1) AS roe_pct,
 ROUND(u.de, 2) AS debt_to_equity,
 ROUND(u.marketCap / 1e9, 2) AS mktcap_b
FROM universe u JOIN sector_medians sm ON u.exchange = sm.exchange AND u.sector = sm.sector
WHERE u.ev_ebitda / sm.median_ev_ebitda < 0.70
ORDER BY u.ev_ebitda / sm.median_ev_ebitda ASC LIMIT 30

Run this screen on Ceta Research

What We Found

24 invested years. 44% win rate. MaxDD worse than the index.

The down-capture of 51.03% means the strategy absorbs about half of SPY's downturns, which is the one genuine bright spot. But the max drawdown of -44.68%, worse than SPY's -34.90%, shows that in Canada's specific crash environment (2008), the portfolio fell harder than the index.

The Sharpe ratio of 0.292 is below SPY's 0.322. On a risk-adjusted basis, this strategy underperforms the index.

What happened in the early years

The strategy's positive CAGR case rests heavily on 2001-2005:

Those early years reflect a period when Canadian resource stocks were cheap relative to their own sector peers, and then the commodity supercycle of 2003-2007 began closing those discounts. Energy, metals, and materials stocks that were unloved at the end of the dot-com era repriced sharply. The signal happened to be pointing exactly at those names.

Year-by-year returns

2008: the worst year

The -44.7% in 2008 drove the max drawdown beyond SPY's -34.3%. Canadian oil, gas, and mining stocks fell catastrophically as commodity prices collapsed. The EV/EBITDA signal had loaded the portfolio into exactly these sectors because they were trading at discounts to their sector peers. When the global cycle reversed and commodity demand collapsed, those discounts deepened rather than closed.

This is the core problem with sector-relative signals in resource-heavy markets: the signal can't distinguish between a temporary mispricing and a structural repricing of the entire sector.

2012-2015: four years of consistent underperformance

After a decent recovery in 2009-2010, the strategy fell behind SPY in every year from 2012 to 2015. Canadian energy and materials stocks spent these years undergoing a prolonged re-rating as US shale production reshaped global oil markets. The signal kept pointing at resource discounts. Those discounts didn't close on any reasonable timeline.

2016: +24.9% excess, the commodity rebound

Oil prices recovered in 2016. Resource stocks that had been deeply discounted for four years finally repriced. This is the strategy working exactly as intended: the long wait for discount closure. But the patience required (four consecutive underperforming years) is difficult to maintain in practice.

Why Canada's resource composition is a structural problem

The sector-relative signal works when discount stocks are cheap because of temporary sentiment, not because of fundamental sector-level repricing. In Canadian energy and materials, the "discount" often reflects:

1. Lower-quality reserves or assets within the same sector
2. Sector-wide commodity price cycles that depress all sector multiples simultaneously
3. Canadian producers structurally trading at a discount to US peers due to infrastructure and pipeline constraints

The signal sees "cheap vs sector median" and buys. But if the sector median itself is being dragged down by the same forces affecting the discount stocks, mean reversion doesn't occur on any useful timeline.

Backtest Methodology

Full methodology: backtests/METHODOLOGY.md

Limitations

MaxDD worse than SPY. The -44.68% max drawdown exceeds SPY's -34.90%. Canada is the only Western exchange in this study where the strategy worsens maximum drawdown vs the benchmark. Resource sector concentration in down cycles explains most of this.

Win rate 44%. The strategy underperforms SPY in more years than it outperforms. Fewer than half of all years generate positive excess returns.

Sector concentration risk. The TSX is heavily weighted toward energy, financials, and materials. Without a sector cap, the portfolio can cluster in resource stocks. A bad commodity cycle (2012-2015, 2018-2020) creates prolonged stretches of underperformance.

Currency effects. Returns are in CAD. The benchmark (SPY) is in USD. CAD/USD fluctuations add noise. The Canadian dollar tends to track oil prices, which means portfolio returns and currency effects can amplify each other in volatile commodity years.

The CAGR case is fragile. Remove the 2001-2005 period and the strategy underperforms SPY. The long-run positive result depends heavily on the commodity supercycle that opened the backtest window.

Conclusion

Sector-relative EV/EBITDA on Canada's TSX delivers 8.42% CAGR over 24 invested years, technically ahead of SPY's 7.64%. The honest read: +0.78% annual excess return while carrying -44.68% max drawdown (worse than the benchmark), a 44% win rate, and a Sharpe ratio below SPY.

The signal works in Canada when sector discounts have a specific catalyst, like the 2003-2005 commodity supercycle or the 2016 oil rebound. Between those windows, it's a waiting game with no guarantee the discount closes on any useful timeline.

For investors in Canada's resource-heavy market, the question isn't whether value investing works in theory. It's whether "cheap vs sector median" is a reliable signal when the sector itself is cyclically distressed. The data suggests it often isn't.

Part of a Series: Global | US | UK | Switzerland | Japan | India | Hong Kong | Germany

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 financial data warehouse). Returns in CAD (local currency). Benchmark S&P 500 in USD. Past performance doesn't guarantee future results. See full methodology at github.com/ceta-research/backtests.