The Empirical Foundation: Momentum as a Persistent Market Anomaly
Trend following is not merely a trading tactic; it is a rigorously documented market anomaly that has persisted across centuries, asset classes, and geographies. The foundational logic rests on the momentum effect—the empirically observed tendency for assets that have performed well recently to continue performing well, and vice versa. This contradicts the weak-form efficient market hypothesis, which posits that past prices cannot predict future returns. Researchers Jegadeesh and Titman, in their seminal 1993 paper “Returns to Buying Winners and Selling Losers,” demonstrated that a strategy of buying past three-to-twelve-month winners and shorting past losers generated significant abnormal returns of approximately 1% per month. This finding has been replicated in over 40 countries and across equities, currencies, commodities, and fixed income. The persistence of momentum across such diverse markets suggests it is not a statistical fluke but a structural feature of how markets process information. Trend following, therefore, is the systematic exploitation of this momentum anomaly using price-based rules, allowing traders to capture the fat tail of market returns that arise from gradual information diffusion and herding behavior.
Behavioral Finance: The Psychological Engine of Trend Persistence
The logic behind trend following is deeply rooted in behavioral finance, specifically the cognitive biases that cause investors to underreact and then overreact to new information. The underreaction hypothesis, supported by Barberis, Shleifer, and Vishny (1998), posits that investors anchor to prior beliefs and update them conservatively, causing prices to adjust gradually to fundamental news. This gradual adjustment creates the initial phase of a trend. As the trend continues and becomes visible, the disposition effect—the tendency to sell winners too early and hold losers too long—creates supply-demand imbalances that further fuel the trend. Subsequently, the overreaction phase, driven by herd behavior and fear of missing out (FOMO), causes prices to overshoot fundamental value. Trend following strategies are designed to capture the meat of these moves by entering after the underreaction phase has begun and exiting before the overreaction fully reverses. This behavioral architecture explains why trends are not random walks; they are the visible signature of human psychology playing out in aggregate through market prices.
The Mathematics of Asymmetric Returns: Cutting Losses and Letting Profits Run
The core mathematical logic of trend following is the deliberate construction of an asymmetric return distribution. By definition, a trend follower enters a position when price breaks above a resistance level (or below support) and exits when price reverses by a predetermined amount or breaches a trailing stop. This creates a payoff profile where losses are capped at a small fraction of capital, while winners are allowed to run for as long as the trend persists. The result is a low win rate—often 35% to 45%—but a high reward-to-risk ratio. For example, a strategy with a 40% win rate and a 3:1 average win/loss ratio yields a positive expectancy of 0.6R per trade. This asymmetry is the mathematical engine of trend following. It does not require predicting the future; it only requires that when trends occur, they are large enough to offset the many small losses incurred during choppy, non-trending markets. This is the same logic that underpins insurance: small, frequent premiums (losses) are paid in exchange for rare but large payouts (wins).
The Convexity Advantage: Positive Skewness and Tail Hedging
Trend following strategies exhibit positive skewness, meaning they generate many small losses and a few large gains. This convexity is a unique attribute that distinguishes trend following from most other investment strategies, which typically exhibit negative skewness (many small gains and rare large losses). The positive skewness of trend following makes it a natural hedge against tail risk. During market crashes—such as 2008, 2020, and 2022—trend following strategies often perform exceptionally well because they are positioned short or out of the market when prices collapse. This was empirically demonstrated in 2008, when the SG Trend Index returned +11.6% while the S&P 500 lost 37%. The logic is straightforward: a trend follower does not need to predict a crash; the price action itself triggers a short position as the trend turns down. This convex payoff profile makes trend following an attractive diversifier for traditional portfolios, as it provides crisis alpha—returns that are uncorrelated or negatively correlated with equities during extreme downturns.
Information Diffusion and the Slow Incorporation of News
Markets are not instantaneous information-processing machines. Information diffuses slowly through the investor base due to differences in access, attention, and interpretation. Institutional investors, retail traders, and algorithmic systems all process the same news at different speeds. This slow diffusion creates serial correlation in returns—the statistical signature of trends. Research by Hong and Stein (1999) formalized this in their “gradual information diffusion” model, showing that when information is costly to acquire and process, prices adjust gradually, generating momentum. Trend following exploits this by systematically entering positions after the initial price move has begun, effectively piggybacking on the delayed reaction of the broader market. The strategy does not require understanding why the trend is happening; it only requires that the trend is happening. This agnosticism to fundamentals is a feature, not a bug: it allows trend followers to profit from trends driven by factors they could never predict, such as geopolitical shocks, technological disruptions, or central bank policy shifts.
The Role of Volatility and Risk Management in Trend Following
Volatility is the raw material of trend following. Without volatility, there are no trends to capture. However, volatility must be managed, not avoided. The logic of trend following incorporates volatility normalization: positions are sized inversely to market volatility, ensuring that each trade risks a consistent fraction of capital regardless of the underlying asset’s noise. This approach, known as volatility targeting, prevents a single high-volatility trade from derailing the portfolio. Furthermore, trend following strategies typically use trailing stops that are a function of volatility (e.g., 2x ATR), allowing them to adapt to changing market conditions. The mathematical logic is that trends are more likely to persist when volatility is moderate and trending, and more likely to whipsaw when volatility is spiking randomly. By dynamically adjusting position size and stop levels, trend followers maintain a consistent risk profile while allowing the asymmetric payoff structure to work over time. This risk management framework is what allows trend following to survive drawdowns and remain in the game long enough to capture the next large trend.
Cross-Asset Diversification: The Portfolio-Level Logic
The logic of trend following is amplified at the portfolio level through cross-asset diversification. Trends occur independently across equities, bonds, currencies, and commodities, driven by different fundamental and behavioral factors. A trend following portfolio holds dozens or hundreds of positions simultaneously, each sized to a consistent risk budget. This diversification reduces the impact of any single trade’s loss while preserving the upside of any single trade’s gain. The mathematics of diversification—the square root of time rule and the reduction of portfolio variance—means that a portfolio of weakly correlated trend following strategies has a much higher risk-adjusted return than any single strategy. This is why managed futures funds, which are predominantly trend following, have historically exhibited low correlation to traditional stock-bond portfolios. The portfolio-level logic is that the sum of many small, independent asymmetric bets produces a smoother equity curve than any individual bet, while the positive skewness of the individual bets is preserved at the portfolio level.
The Adaptive Nature of Trend Following: No Fixed Parameters
A common misconception is that trend following is a static set of rules. In reality, the logic of trend following is adaptive. The strategy does not assume that trends will always last a certain number of days or that a specific moving average will always work. Instead, it uses a range of parameters—short-term, medium-term, and long-term trend signals—to capture trends of different durations. This parameter diversity is a form of ensemble learning: by combining multiple trend signals, the strategy reduces the risk of overfitting to any single historical period. Furthermore, trend following adapts to changing market regimes. In a low-volatility, trending market, the strategy will be fully invested; in a high-volatility, mean-reverting market, it will be mostly in cash or hedged. This adaptability is the logical response to the fact that markets are non-stationary. A strategy that works in one decade may fail in the next, but a strategy that adapts to price action itself—rather than to a fixed set of fundamental assumptions—has a structural edge that persists across regimes.
Empirical Performance Across Centuries and Crises
The empirical record of trend following is not limited to the modern era. Research by Hurst, Ooi, and Pedersen (2017) in “A Century of Evidence on Trend-Following Investing” examined trend following across equities, bonds, currencies, and commodities from 1880 to 2016. They found that a diversified trend following strategy delivered an annualized gross Sharpe ratio of 0.76, with positive returns in every decade except the 1930s. More importantly, the strategy performed well during eight of the ten largest equity market drawdowns in the century, including the Great Depression, the 1970s stagflation, the 2000 dot-com crash, and the 2008 financial crisis. This empirical persistence is the ultimate validation of the strategy’s logic: it is not dependent on a specific market environment, a specific central bank policy, or a specific technology. It is dependent only on the existence of trends, which are a permanent feature of markets driven by human psychology and slow information diffusion.
The Execution Edge: Systematic Discipline Over Discretionary Bias
Finally, the logic of trend following includes a meta-logical advantage: it is systematic. By codifying entry, exit, and position sizing rules, trend following eliminates the discretionary biases that destroy most traders’ returns. Fear, greed, regret, and overconfidence are replaced by algorithms that execute the same logic regardless of recent performance. This systematic discipline is not a minor detail; it is the mechanism that ensures the asymmetric payoff structure is actually realized. A discretionary trader might cut a winning trend short due to fear or hold a losing position due to hope, destroying the very asymmetry that makes trend following work. The systematic nature of trend following enforces the discipline required to let profits run and cut losses short. In this sense, the logic of trend following is not just about market behavior; it is about exploiting the predictable failures of human psychology—including the psychology of the trader themselves. By removing the human element from execution, trend following ensures that the strategy’s mathematical edge is captured consistently over time.
