Exponential Moving Average
An Exponential Moving Average (EMA) is a type of moving average that assigns exponentially decreasing weights to historical prices, with the most recent prices receiving greater weight than older prices, making the EMA more responsive to recent market developments than a simple moving average (SMA) of the same length. EMAs are widely used to identify trend direction, generate trading signals through crossovers, and construct more complex momentum indicators.
Key takeaways
- The EMA weighting factor k = 2/(n+1), where n is the period; a 20-day EMA assigns k = 2/21 ≈ 9.5% weight to today's price.
- Short-period EMAs (5-day, 9-day) are highly responsive to price changes and generate more trading signals, while longer-period EMAs (50-day, 200-day) provide smoother trend identification.
- The 'golden cross' (short EMA crossing above long EMA) is a bullish signal; the 'death cross' (short crossing below long) is bearish.
- MACD (Moving Average Convergence Divergence)—one of the most popular technical indicators—is calculated as the difference between a 12-day and 26-day EMA.
- EMAs are the foundation of many systematic trading strategies, particularly trend-following CTAs that trade across multiple asset classes.
Explanation
The exponential moving average improves upon the simple moving average by addressing its two main weaknesses: the SMA assigns equal weight to all observations within the window (treating a 200-day-old price as equally informative as yesterday's price) and it exhibits 'ghost effects' when old data drops off the window. The EMA, by contrast, never fully discards old data—it smoothly decays the influence of historical observations at an exponential rate, making current prices more influential while preserving the context of historical trends.
The EMA is computed recursively: EMA_t = Price_t × k + EMA_{t-1} × (1-k), where k = 2/(n+1) and n is the specified period. This recursive formula means the EMA can be updated with only the current price and the previous EMA value—no need to store a full window of historical prices. The half-life of the EMA (the time until an observation's weight decays to half its initial value) is approximately n/1.44, providing intuition for the effective memory of each EMA period: a 10-day EMA has a half-life of approximately 7 days.
Trend identification is the primary application of EMAs in technical analysis. A price trading above its 200-day EMA is interpreted as being in an uptrend; below is interpreted as a downtrend. The 50-day EMA crossing above the 200-day EMA (the 'golden cross') is one of the most widely watched technical signals in equity markets, generating coverage in financial media and often creating self-fulfilling momentum as retail investors act on the signal. Academic research on moving average crossover signals finds modest but statistically significant predictive power in certain markets and time periods, though transaction costs often eliminate the paper profit for high-frequency implementations.
The MACD indicator—Moving Average Convergence/Divergence—extends EMA analysis by measuring the difference between two EMAs (typically 12-day and 26-day). When the shorter EMA is above the longer (MACD is positive), the trend is interpreted as bullish; when below, bearish. A 9-day EMA of the MACD is called the 'signal line,' and crossovers between MACD and signal line generate buy/sell signals. MACD divergence—when price makes new highs or lows but MACD does not confirm—is interpreted as a trend exhaustion warning.
For systematic hedge funds and CTAs (Commodity Trading Advisors), EMAs are foundational building blocks of trend-following strategies. A simple trend-following model might buy an asset when its price is above the 100-day EMA and sell when below, applied across 50–100 liquid futures markets simultaneously. The diversification across uncorrelated markets with independent trends is why CTA returns have historically shown low correlation to equity and bond markets—particularly valuable during equity bear markets when CTA trend-following strategies often generate their best absolute returns.
Formula
EMA_t = Price_t × k + EMA_{t-1} × (1 - k); k = 2 / (n + 1); MACD = EMA(12) - EMA(26)
Example
A stock closes at the following prices over 5 days: $100, $102, $105, $103, $107. Computing the 3-day EMA (k = 2/(3+1) = 0.5): Starting EMA = Day 1 close = $100. Day 2: EMA = 102 × 0.5 + 100 × 0.5 = $101.00. Day 3: EMA = 105 × 0.5 + 101 × 0.5 = $103.00. Day 4: EMA = 103 × 0.5 + 103 × 0.5 = $103.00. Day 5: EMA = 107 × 0.5 + 103 × 0.5 = $105.00. For comparison, the 3-day simple moving average on Day 5 = (103 + 105 + 107)/3 = $105.00—the same in this case, but in general the EMA responds faster to trend changes and would diverge from the SMA following sharp reversals.
Related terms
Bollinger Bands Bond Convergence Correlation Diversification Equity Head And Shoulders Pattern Momentum Indicator Moving Average On Balance Volume Overbought Paper Profit