A relative strength model ranks securities by comparing their recent performance to each other, then selects the strongest performers for inclusion in a portfolio. Unlike absolute return measures, relative strength asks not "did this stock go up?" but "did this stock outperform its peers?" The distinction matters because leadership within a group tends to persist for months at a time.

The model builds on a well-documented empirical finding: securities that have outperformed their peers over the past 6 to 12 months tend to continue outperforming, while the weakest tend to keep lagging. This persistence effect was first documented by Robert Levy in 1967 and later confirmed by extensive academic research, most notably Jegadeesh and Titman's 1993 study of momentum returns. Relative strength is closely related to momentum but focuses specifically on cross-sectional ranking rather than absolute price trends.

Conceptual Framework

Relative strength rests on the same behavioral and structural forces that drive momentum: investor underreaction to new information, gradual diffusion of news across market participants, and institutional herding. When a stock begins outperforming its peers, the factors driving that outperformance (strong earnings growth, favorable industry trends, institutional accumulation) tend to persist for months rather than reversing immediately.

The approach differs from absolute momentum or trend-following models in one critical way. A trend-following model asks whether a security is in an uptrend or downtrend. Relative strength asks which securities are leading and which are lagging within a defined peer group. During a broad market decline, every stock may be falling, but relative strength still identifies which ones are declining least, which can be valuable for defensive positioning.

Absolute vs. Relative Performance

An important distinction separates absolute and relative performance measures:

• Absolute performance: The total return of a security over a period, measured in isolation. A stock that returned 15% over the past year has strong absolute performance regardless of what the broader market did. Absolute measures can identify securities in uptrends but do not indicate whether those securities are leading or lagging their peers.
• Relative performance: The return of a security compared to a reference point, typically a benchmark index or the median return of a peer group. A stock that returned 15% has weak relative strength if the benchmark returned 25%, and strong relative strength if the benchmark returned 5%. Relative measures identify leadership within a group.
• Combining both: Many practitioners use both dimensions together. A security with strong relative strength and positive absolute performance is in a leadership position within a rising market. A security with strong relative strength but negative absolute returns is simply declining less than its peers, a weaker signal.

It is important not to confuse relative strength with the Relative Strength Index (RSI), a popular technical indicator created by J. Welles Wilder. RSI is a momentum oscillator that measures a single security's recent gains against its recent losses on a scale of 0 to 100. Relative strength, by contrast, compares different securities to each other. They are entirely different concepts despite the similar names.

Core Assumptions

Relative strength models make several assumptions about how cross-sectional performance patterns behave:

• Performance persistence: The model assumes that relative rankings are somewhat stable over time. This is well-supported by academic evidence for holding periods of 1 to 12 months, but the effect weakens at very short (under one month) and very long (over one year) horizons.
• Rankings contain information: The model assumes that the spread in performance across securities is meaningful, not random noise. If all securities in the universe are moving in lockstep, the ranking carries little signal. Relative strength works best when there is meaningful dispersion (variation in returns) across the universe.
• The universe is appropriate: The choice of which securities to rank against each other matters. Comparing a small-cap biotech stock to a large-cap utility within the same ranking produces a signal that reflects sector and size differences more than genuine stock selection.
• Lookback period captures the relevant trend: The lookback period must be long enough to capture a meaningful trend but short enough to remain responsive. Too short a window produces noisy rankings that whipsaw. Too long a window produces stable rankings that react slowly to genuine changes in leadership.

Signal Construction

A relative strength model follows a five-step pipeline from defining the comparison universe to assembling the final portfolio signal.

Universe Definition

The universe should contain securities that are genuinely comparable: similar in market capitalization range, liquidity, and investment character. Common universe definitions include all constituents of a broad index (such as the S&P 500), all stocks within a specific sector, or a curated list of ETFs representing different asset classes.

Liquidity filters are important. Securities with very low trading volume can produce misleading relative strength readings because their prices may not update as frequently, creating stale rankings. Minimum average daily volume thresholds, minimum market capitalization requirements, and exclusions for recent IPOs (initial public offerings) help ensure that the rankings reflect genuine performance differences.

Return Calculation

For each security in the universe, the model calculates the total return over one or more lookback periods. The most common specification uses total return (price change plus dividends) over the trailing 6 or 12 months. As with momentum models, many implementations skip the most recent month to avoid contamination from short-term reversal effects, where very recent winners tend to briefly reverse.

Some implementations calculate the return relative to a benchmark rather than in absolute terms. The model computes the excess return of each security over the benchmark index, then ranks securities by that excess return. This isolates the stock-specific component of performance from the market-wide movement.

Cross-Sectional Ranking

After calculating returns, the model ranks all securities from strongest to weakest. The raw ranking is typically converted to a percentile score so that the output is consistent regardless of universe size. A percentile rank of 95 means the security outperformed 95% of its peers over the lookback period.

Composite relative strength scores combine rankings from multiple lookback periods into a single number. A common approach assigns weights to short-term (1 to 3 months), medium-term (3 to 6 months), and long-term (6 to 12 months) relative strength, then blends them. For example, a composite might weight the 3-month return at 40%, the 6-month return at 30%, and the 12-month return at 30%. This multi-timeframe approach captures both recent acceleration and sustained leadership.

Signal Generation

The composite ranking produces the selection signal. Securities in the top quintile (top 20%) or top decile (top 10%) are candidates for purchase. The threshold between "buy" and "hold" zones determines how concentrated the portfolio will be.

Many implementations add hysteresis (a buffer zone) to reduce unnecessary trading. A security might need to rank in the top 20% to enter the portfolio but would not be sold unless it drops below the top 40%. This buffer prevents churning when a security's rank oscillates near the threshold.

Risk Architecture

Relative strength strategies share many risk properties with momentum strategies, because both rely on the persistence of recent performance. The cross-sectional ranking approach introduces some additional considerations.

Regime Sensitivity

Relative strength models perform best during sustained trending environments where clear performance leaders emerge and maintain their advantage. During choppy, range-bound markets where leadership rotates quickly, the model generates frequent ranking changes that can lead to whipsaw trades: buying a security just as its relative strength fades, then selling it as it re-establishes leadership.

The transition between market regimes is particularly hazardous. When the market shifts from a growth-led rally to a value-led recovery (or vice versa), the entire ranking can invert. Securities that ranked at the top during one regime can drop to the bottom in the new regime, and the model may be slow to recognize the change, especially with longer lookback windows.

Crowding and Momentum Crash Risk

Because relative strength and momentum signals are closely related, they tend to identify similar securities. When many institutional investors, ETFs, and systematic strategies hold the same top-ranked positions, the resulting crowding amplifies the risk of a coordinated unwinding. If a negative shock hits the top-ranked group, forced selling by multiple holders at once can produce losses far larger than the underlying fundamentals would justify.

Known Limitations

Practical Considerations

Multi-Timeframe Analysis

Combining relative strength across multiple timeframes produces a more robust signal than any single lookback period. Short-term relative strength (1 to 3 months) captures recent acceleration. Medium-term (3 to 6 months) confirms that the trend has substance. Long-term (6 to 12 months) identifies sustained outperformers with durable advantages.

Securities that rank highly across all three timeframes are the strongest candidates: they are leading peers over the long term, have maintained that leadership through the medium term, and are continuing to accelerate in the short term. Securities that show short-term improvement but poor long-term rankings may be experiencing a temporary bounce rather than genuine trend establishment.

Rebalancing Frequency

Relative strength portfolios are typically rebalanced monthly or quarterly. Monthly rebalancing keeps the portfolio aligned with the latest rankings but generates more trading. Quarterly rebalancing reduces costs but allows positions to drift further from the target signal. The appropriate frequency depends on the lookback periods used and the liquidity of the universe.

Universe and Sector Considerations

Relative strength can be applied at multiple levels. At the asset class level, it compares broad market segments (domestic stocks, international stocks, bonds, commodities) to identify which asset classes are leading. At the sector level, it compares industry groups within the equity market. At the stock level, it ranks securities within a sector or across the full market.

A common two-stage approach first uses relative strength to identify the strongest sectors, then applies a second screen within those sectors to select individual stocks. This top-down approach aligns the portfolio with prevailing sector trends while selecting the best opportunities within each sector, and helps manage concentration risk by controlling sector weights.

Related Models

Further Reading

• Jegadeesh, N. and Titman, S. (1993). "Returns to Buying Winners and Selling Losers: Implications for Stock Market Efficiency." The Journal of Finance, 48(1), 65–91.
• Levy, R.A. (1967). "Relative Strength as a Criterion for Investment Selection." The Journal of Finance, 22(4), 595–610.
• Asness, C.S., Moskowitz, T.J., and Pedersen, L.H. (2013). "Value and Momentum Everywhere." The Journal of Finance, 68(3), 929–985.
• Daniel, K. and Moskowitz, T.J. (2016). "Momentum Crashes." Journal of Financial Economics, 122(2), 221–247.
• O'Shaughnessy, J.P. (2011). What Works on Wall Street: The Classic Guide to the Best-Performing Investment Strategies of All Time. McGraw-Hill (4th ed.).
• Faber, M.T. (2007). "A Quantitative Approach to Tactical Asset Allocation." The Journal of Wealth Management, 10(1), 69–79.
• "Momentum and Contrarian Strategies" (CFA Institute Professional Learning).