SystemTrader
ETFsAdvancedFactor-Based

Alpha Rotation

A sophisticated rotation strategy that selects sector ETFs based on their Fama-French alpha—the portion of returns unexplained by common risk factors. Unlike raw momentum, alpha rotation identifies true outperformance after adjusting for market, size, and value exposures.

Universe
10-11 Sector ETFs
Lookback
36 Months
Holding
1 Month
Selection
Positive Alpha

Overview

The alpha rotation strategy improves upon simple momentum by using risk-adjusted alpha instead of raw returns. Alpha is the intercept from regressing sector returns against the Fama-French factors (market, size, value, profitability, investment), representing returns that cannot be explained by exposure to common risk factors.

The intuition is simple: a sector showing high raw returns might just be taking on more market risk. By computing alpha, we isolate the true excess return—the portion that represents genuine outperformance rather than factor loading. Positive alpha indicates skill or an anomaly worth exploiting.

Research by Sarwar, Mateus, and Todorovic (2017) found that a long-only sector rotation strategy buying sectors with positive five-factor alpha generates four times higher Sharpe ratio than S&P 500 buy-and-hold. When combined with a recession filter, the Sharpe ratio improvement is ten-fold.

Key Insight

Factor-Adjusted
Strips out market/size/value effects
4x Sharpe Ratio
vs S&P 500 buy-and-hold
True Outperformance
Identifies genuine alpha

Sector Alpha Rankings

Sectors ranked by Fama-French 3-factor alpha.Sectors with positive alpha (α > 0) are selected for the portfolio.

Selected Sectors
5 of 10
Positive alpha
Average Alpha
+2.2%
Selected sectors
Regression Window
36 Mo
Rolling lookback
Positive alpha (Buy)
Negative alpha (Avoid)
Alpha = 0 threshold

Research

Alpha-based rotation combines insights from factor investing and sector rotation research. By using regression alpha instead of raw returns, the strategy identifies sectors with genuine excess returns after controlling for systematic risk exposures.

Factor Momentum Research

The Mathematics

In Plain English

The math behind this strategy is straightforward. Here's what you're actually doing:

  1. 1
    Collect 36 months of monthly returns for each sector ETF
  2. 2
    Run a regression of each sector's returns against the Fama-French factors
  3. 3
    Extract the alpha (intercept) from each regression
  4. 4
    Select all sectors with positive alpha for the portfolio
  5. 5
    Equal-weight the selected sectors and hold for one month
  6. 6
    Repeat monthly with updated rolling regressions

That's it. The formulas below just express this process precisely.

Technical Formulas

1
Fama-French Three-Factor Regression

Formula
R_i(t) - R_f(t) = \alpha_i + \beta_{1,i} \cdot MKT(t) + \beta_{2,i} \cdot SMB(t) + \beta_{3,i} \cdot HML(t) + \epsilon_i(t)

Where R_i(t) is sector i return, R_f is risk-free rate, MKT is market excess return, SMB is small-minus-big (size), HML is high-minus-low (value). Alpha (α_i) is the unexplained excess return.

2
Fama-French Five-Factor Regression

Formula
R_i(t) - R_f(t) = \alpha_i + \beta_1 MKT + \beta_2 SMB + \beta_3 HML + \beta_4 RMW + \beta_5 CMA + \epsilon_i(t)

Adds RMW (robust-minus-weak profitability) and CMA (conservative-minus-aggressive investment). The five-factor model better fits sector returns and produces more reliable alpha estimates.

3
Selection Rule

Formula
Portfolio = \{ ETF_i : \alpha_i > 0 \}

Select all sector ETFs with positive estimated alpha. Equal-weight the selected sectors. If no sectors have positive alpha, hold cash or a broad market ETF.

Rolling WindowNote

Use a 36-month rolling window to estimate alpha. This balances stability (longer window) with responsiveness (shorter window). Some implementations use 60 months for more stable estimates.

Factor Data SourceNote

Fama-French factor returns are available free from Kenneth French's data library. Use monthly factor returns matching your ETF return frequency.

Strategy Rules

Universe & Data

  1. Use the 10 S&P 500 sector ETFs (XLK, XLV, XLF, XLY, XLP, XLE, XLI, XLB, XLRE, XLU)
  2. Collect at least 36 months of monthly return history
  3. Download Fama-French factor data from Kenneth French's website
  4. Use total returns (dividend-adjusted) for sector ETFs
  5. Match the frequency: monthly returns with monthly factors

Alpha Calculation

  1. 1Run OLS regression of excess returns on FF3 or FF5 factors
  2. 2Use 36-month rolling window ending at month-end
  3. 3Extract the intercept (alpha) from each regression
  4. 4Record statistical significance (t-stat) if filtering by it
  5. 5Update regressions monthly with new data

Portfolio Construction

  1. Select all sectors with alpha > 0
  2. Equal-weight selected sectors
  3. If no positive alphas, hold cash or broad market ETF (SPY)
  4. Optional: require t-stat > 1.5 for stronger conviction
  5. Rebalance on the last trading day of each month

Risk Management

  1. 1Monitor number of selected sectors (diversification)
  2. 2Consider recession filter: hold bonds during NBER recessions
  3. 3Set maximum sector allocation (e.g., 25% per sector)
  4. 4Track factor exposures of overall portfolio
  5. 5Review regression fit (R-squared) for reliability

Implementation Guide

Implementing alpha rotation requires running monthly regressions, which is more complex than simple momentum but provides risk-adjusted signals. Python or R makes this straightforward.

1

Gather Return Data

Collect monthly total returns for the 10 sector ETFs going back at least 36 months. Use adjusted close prices to capture dividends.

Tips
  • Data sources: Yahoo Finance, Alpha Vantage, Tiingo
  • Calculate monthly returns: (P_t / P_{t-1}) - 1
  • Align dates to end-of-month
  • Store in a structured format (CSV, DataFrame)
2

Download Fama-French Factors

Get the monthly Fama-French factor returns from Kenneth French's data library. These are freely available and updated regularly.

Tips
  • URL: mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html
  • Download "Fama/French 3 Factors" or "Fama/French 5 Factors (2x3)"
  • Factors are in percentage form—match your ETF return format
  • RF (risk-free rate) is included in the file
3

Run Rolling Regressions

For each sector, run a regression of excess returns (R_i - RF) on the factors using a 36-month rolling window. Extract the alpha (intercept) from each regression.

Tips
  • Python: use statsmodels OLS or sklearn LinearRegression
  • R: use lm() function
  • Excess return = ETF return - RF
  • Save alpha, t-statistic, and R-squared for each sector

Ensure your factor data dates align exactly with your ETF return dates. Misalignment causes incorrect alpha estimates.

4

Select Positive Alpha Sectors

At month-end, identify all sectors with positive alpha. These are your buy candidates. Equal-weight them for the next month's portfolio.

Tips
  • Simple rule: alpha > 0 means include
  • Stricter rule: require t-stat > 1.5 or 2.0
  • If no positive alphas, hold cash or SPY
  • Record your selections for performance tracking
5

Execute & Rebalance Monthly

On the last trading day of each month, rebalance to the new set of positive-alpha sectors. Sell sectors that turned negative, buy new positive-alpha sectors.

Tips
  • Use limit orders around the close
  • Equal-weight means 100% / number of selected sectors
  • Typical selection: 4-7 sectors out of 10
  • Track turnover for tax efficiency

Tools Required

This strategy requires running regressions, so you'll need Python, R, or Excel with regression capability. A basic broker account is sufficient for execution. Consider automating the monthly calculation with a script.

Helpful Tools & Resources

Factor Data
Kenneth French Data Library (free)
Regression Analysis
Python (statsmodels), R, Excel
Execution
Schwab, Fidelity, Interactive Brokers

Strategy Variations

Explore different ways to implement this strategy, each with its own trade-offs and benefits.

Five-Factor Alpha

Use the Fama-French five-factor model instead of three factors. Adds profitability (RMW) and investment (CMA) factors for more precise alpha estimates.

Research shows five-factor model fits sector returns better.

Recession Filter

During NBER-defined recessions, move entirely to bonds or cash regardless of sector alphas. Dramatically improves Sharpe ratio by avoiding drawdowns.

Ten-fold Sharpe improvement vs. buy-and-hold.

Alpha Momentum

Instead of just positive/negative alpha, rank sectors by alpha and select top N. Combines alpha signal with momentum-style ranking.

More concentrated portfolio, potentially higher returns.

Long/Short Alpha

Go long sectors with positive alpha and short sectors with negative alpha. Creates a market-neutral portfolio betting on alpha dispersion.

Requires margin and short-selling capability.

Statistical Significance Filter

Only select sectors where alpha has t-statistic > 2.0, indicating statistical significance. Reduces false positives from noisy estimates.

Fewer positions but higher conviction.

Consider combining multiple variations or testing them against your specific investment goals and risk tolerance.

Risks & Limitations

High(2)
Medium(4)
Alpha Estimation ErrorHigh

Alpha is estimated from a 36-month regression and contains estimation error. Past alpha may not persist, especially when based on noisy data or short windows.

Impact:
Factor Regime ChangesHigh

The relationship between sectors and factors can change over time. A sector's alpha may flip from positive to negative as market regimes shift.

Impact:
Model MisspecificationMedium

The Fama-French model may not capture all relevant risk factors. Apparent alpha could actually be compensation for omitted risks.

Impact:
Data and Execution ComplexityMedium

Running monthly regressions requires data management and coding skills. Errors in factor alignment or return calculation can invalidate the strategy.

Impact:
Sector ConcentrationMedium

When few sectors have positive alpha, the portfolio becomes concentrated. This increases idiosyncratic risk and volatility.

Impact:
Declining Alpha Over TimeMedium

As factor investing becomes mainstream, sector alphas may shrink. More capital chasing the same signals can arbitrage away excess returns.

Impact:
Understanding these risks is essential for proper position sizing and portfolio construction. Consider combining with other strategies to mitigate individual risk factors.

References

  • Sarwar, G., Mateus, C., & Todorovic, N. (2017). US Sector Rotation with Five-Factor Fama-French Alphas. Journal of Asset Management, 19(2), 116-132 [Link]
  • Hübner, G. (2018). Alpha Momentum and Price Momentum. International Journal of Financial Studies, 6(2), 49 [Link]
  • Fama, E.F. & French, K.R. (2015). A Five-Factor Asset Pricing Model. Journal of Financial Economics, 116(1), 1-22 [Link]
  • Ehsani, S. & Linnainmaa, J. (2022). Factor Momentum and the Momentum Factor. Journal of Finance, 77(3), 1877-1919 [Link]

ETF trading involves risk of loss. Alpha rotation requires running statistical regressions and may produce unreliable signals during regime changes. Past alpha does not guarantee future alpha. This is educational content, not investment advice.

Ready to explore more ETF strategies?

Discover additional tactical ETF strategies including sector momentum, factor timing, and risk parity approaches.

Browse ETF Strategies