Why smart beta underdelivers its backtest

What a factor is

A factor is a shared characteristic that has explained differences in returns over long periods: cheap stocks (value), recent winners (momentum), stable earners (quality), calmer stocks (low volatility), small companies (size).

Funds that screen on those characteristics are called factor ETFs, or smart beta. Nobody picks stocks by hand as in an active fund, but they do not simply follow market cap order either.

The academic grounding is real. Across decades of data, these characteristics genuinely produced return differences.

The gap between backtest and live

Paper returns typically exclude trading costs or assume very low ones. But factor strategies must refresh their holdings on a schedule, and that turnover carries commissions and market impact.

Momentum is the clearest case. It has to keep chasing what has already risen, so turnover is high, and realistic cost assumptions erode a large share of the paper edge.

On top of that sits the fund's expense ratio. Factor funds cost more than broad index funds, and the difference comes straight out of the expected excess return.

Publication shrinks the edge

Once an anomaly is published, capital arrives to exploit it. Studies have found that documented excess returns tend to decline noticeably after publication.

So much of any backtest window covers a period when nobody knew yet. The product reaches shelves after that period has ended.

And factors can underperform for years — sometimes more than a decade. 'Works over the long run' and 'works while I hold it' are entirely different claims. Most investors sell somewhere inside the drought.

Same label, different definition

Not all 'value' funds are alike. One screens on price-to-book, another on earnings or cash flow multiples, another blends several metrics. Different criteria produce substantially different holdings.

'Quality' is vaguer still. It can mean return on equity, leverage, or earnings stability. The same label can wrap completely different portfolios.

So choose a factor fund by its methodology document, not its name. Which metrics, in what order, screened how often — that is the actual product.

None of this argues against factors. It argues against treating a backtested number as an expected return. The question is whether a reason survives after costs and after you account for the drought years.