SMA Trend Filter: The Moving Average Gate for Downtrend Protection

The SMA trend filter is a binary protection mechanism that compares an asset's current price against its simple moving average over a specified period — most commonly ten or thirteen months. If the price is above the moving average, the asset passes the filter and remains eligible for the portfolio. If below, the asset fails and its allocation moves to cash or a defensive alternative. The filter does not rank assets or determine how much to hold — it simply answers a yes-or-no question: is this asset in an uptrend?

How It Works

The simple moving average is the arithmetic mean of an asset's closing prices over the lookback period. A ten-month SMA averages the last ten monthly closing prices. When the current price crosses above this average, it indicates that recent prices are higher than the recent average — a mathematical confirmation that an uptrend is underway. When the price crosses below, recent prices are lower than the average, indicating a downtrend.

The filter operates independently per asset. In a multi-asset portfolio, each position is evaluated against its own moving average. This independence means the portfolio can be partially invested and partially in cash simultaneously — a feature that distinguishes SMA-filtered strategies from binary all-or-nothing approaches. During a sector-specific decline, only the affected position moves to cash while others remain invested. During a broad market crisis, multiple positions may fail simultaneously, automatically increasing the portfolio's cash allocation.

Why Moving Averages Work as Filters

The effectiveness of moving average filters rests on the empirical observation that asset class trends tend to persist for months or years. When an asset's price falls below its long-term average, it is more likely to continue declining than to immediately reverse. This persistence is driven by the same behavioral and structural forces that create momentum: institutional capital flows adjust slowly, macroeconomic conditions evolve gradually, and investor sentiment shifts through stages of denial, recognition, and capitulation that play out over months rather than days.

Meb Faber's landmark research (SSRN #962461) demonstrated that applying a ten-month SMA filter to a diversified portfolio reduced maximum drawdowns by approximately 50% relative to buy-and-hold, with minimal impact on long-term compound returns. The finding has been replicated across asset classes, time periods, and geographies, establishing the SMA filter as one of the most robust and well-validated protective mechanisms in systematic investing.

Strategies That Use SMA Trend Filters

• Ivy Portfolio (GTAA5) — 10-month SMA per asset, 5 positions
• GTAA13, GTAA-AGG3, GTAA-AGG6 — 10-month SMA eligibility gate before momentum ranking
• Risk Parity Trend — US (RP-US) and RP-GLOBAL — SMA filter on inverse-volatility weighted positions
• Trinity Portfolio — SMA filter on the tactical half only
• Tactical Permanent Portfolio (TPP) — SMA filter on equity position only

Limitations

The SMA filter introduces a structural lag: it cannot detect a trend change until the price has moved sufficiently to cross the moving average, which may take one to three months after the actual peak or trough. During sharp, V-shaped corrections like the COVID crash of March 2020, the filter may trigger an exit near the bottom and a re-entry only after a significant portion of the recovery has already occurred. This whipsaw cost is the insurance premium paid for protection during sustained bear markets, where the filter's early exits provide substantial value.