Lookback Period and Turnover: How Parameter Choices Affect Your Portfolio

The lookback period is the number of months of trailing return data a strategy uses to generate its momentum signal. It is arguably the single most important parameter in any tactical allocation strategy — it determines how quickly the strategy detects trend changes, how frequently it trades, and how sensitive it is to market noise versus genuine trend shifts. Understanding the trade-offs inherent in this choice is essential for evaluating whether a strategy's behavior will match your expectations and tolerance for turnover.

The Speed-Stability Spectrum

Shorter lookbacks (1-6 months) produce faster signals that detect trend changes sooner but generate more false alarms and higher turnover. A one-month signal can detect a reversal within a single month but will also react to every significant monthly move, producing frequent position changes that may or may not reflect genuine trend shifts.

Longer lookbacks (10-12 months) produce slower signals that filter out noise but delay the detection of genuine reversals by several months. A twelve-month signal will not turn negative until the asset has been declining long enough for the last year of returns to turn negative — by which point the portfolio may have absorbed significant losses from the initial phase of the decline.

The momentum literature suggests the sweet spot for cross-asset momentum is between three and twelve months, with six months often cited as the horizon where the momentum premium is strongest. This finding has influenced the design of multi-period composites that blend several lookback periods to capture the full spectrum of momentum information.

Common Lookback Periods on the Platform

• 1/3/6 month blend — used by ADM for fastest detection
• 4 months — used by FAA for responsive multi-factor ranking
• 6 months — used by AAA for momentum screening before optimization
• 10 months — used by all Faber GTAA strategies for SMA trend filter
• 12 months — used by GEM and most Keller strategies for signal stability
• 13 months — used by BAA-B for slow asset selection via SMA ratio

Turnover Implications

Shorter lookbacks produce higher turnover. GEM with its twelve-month lookback generates 1-3 trades per year. ADM with its accelerated composite generates 6-10 trades per year. VAA-G4 with its any-one-negative trigger on 13612W generates very high turnover.

Turnover directly affects after-tax returns in taxable accounts. Each trade may trigger short-term capital gains (taxed at ordinary income rates) or long-term gains (taxed at preferential rates). Strategies with annual turnover below 50% are generally considered tax-efficient; those exceeding 200% generate significant annual tax drag that must be weighed against the strategy's pre-tax performance.

Choosing Your Lookback

There is no objectively "best" lookback period — the choice reflects the investor's preference along the speed-stability trade-off. Investors who prioritize drawdown reduction and can tolerate higher turnover should favor shorter lookbacks. Investors who prioritize tax efficiency and behavioral simplicity should favor longer lookbacks. The platform's Scenarios and Builder tools allow investors to test different lookback periods on any strategy, observing how the trade-off plays out across historical market conditions.