Smart Beta and Factor Investing

Where factors come from

The intellectual lineage matters because it defines the bet. Fama and French (1992–93) showed that value and size characteristics explained returns beyond market beta; momentum followed (Jegadeesh–Titman 1993); profitability/quality and low-volatility rounded out the modern canon. Two rival explanations, with different implications you must choose between:

• Risk-based: the premium is payment for bearing something unpleasant (value stocks are distressed-economy-sensitive; small caps are fragile). If true: the premium should persist, because it hurts at the worst times: that's the deal.
• Behavioral: the premium is other investors' systematic mistakes (overpaying for glamour, underreacting to news). If true: it can be arbitraged away as everyone learns; and "everyone" now includes trillions of factor-aware capital.

The uncomfortable synthesis the evidence supports: some of both, plus a third ingredient the replication literature insists on: publication and productization shrink premia (documented post-publication decay across factors). The factor you can buy in an ETF today is, by construction, the famous version of itself.

The record, told honestly in three layers

Layer 1: the long backtests. Multi-decade factor indices show meaningful annualized premia (historically on the order of 1–4% for the canon, before costs). This layer is real and is what every factor brochure shows.

Layer 2: the live-fund era. Since smart beta scaled (roughly 2010s onward): value endured its worst drought on record before the 2021–22 snapback; small-cap's premium has been debated to a whisper in US large-liquid markets; momentum performed but with violent crashes (2009-style reversals are its signature risk); low-vol delivered smoother rides until March 2020 hit its crowded trades hard; quality has been the quiet respectable one. Multifactor funds, sold as the diversified solution, mostly delivered tracking-different mediocrity net of fees. SPIVA-style comparisons of factor funds against plain benchmarks read like the active tables they secretly are.

Layer 3: the investor experience. Factor returns arrive with the worst possible ergonomics: long famines punctuated by fast feasts, which flow data confirms investors reliably miss: buying after feasts, capitulating in famines (Chapter 30's Mistake 1, wearing a lab coat). The behavior gap in factor funds exceeds that in plain index funds: the product's statistical case is strongest exactly where its psychological case is weakest.

Costs, the quiet layer under all three: fees of 0.15–0.40% (UCITS often higher), plus Chapter 41's turnover toll (momentum can turn over 100%+/yr), plus crowding effects around predictable rebalances. "Backtest minus implementation" routinely costs 0.5–1%+ against the paper index: enough to consume a shrunken premium whole.

The buyer's protocol

Factor investing has enough evidence behind it to deserve a standing protocol rather than a warning label. If you tilt, tilt like this:

1. Pick a persistence story you actually believe (risk-based stories age better than behavioral ones, by their own logic) and write it down: it is your only anchor in year six of the famine.
2. Buy the implementation, not the label: methodology (Chapter 40's caps/buffers matter double here), turnover, and above all live multi-year TD versus the plain cap-weighted benchmark: the factor fund must beat the boring fund after everything, or it's a story with a fee.
3. Size for the famine: satellite caps (Chapter 61), sized so a lost decade versus the core is annoying, not plan-breaking. History says a lost decade is not the tail case: it's the recurring case.
4. Never time factors. Factor rotation strategies are the active-management problem squared; the evidence on timing them is dismal even by timing's standards.
5. Accept the null gracefully: the cap-weighted core already holds every factor at market weight; declining this entire aisle is not a knowledge failure: it's the evidence's default.

A worked example

Elif, convinced by the risk-story for value, allocates under protocol:

• The write-up (2019): "Value premia are compensation for cyclical distress exposure; I expect underperformance in growth-led markets and commit through it. Size: 10% satellite. Benchmark: MSCI World. Review: annually; exit only on thesis change, not on drawdown."
• The implementation check: a value UCITS fund whose live 5-yr TD versus its own factor index is tight (craft ✓), fee 0.25%, turnover moderate.
• The famine (2019–2020): the satellite trails the world by 18 points. The written thesis says: expected. She holds: the document, not her mood, is doing the holding.
• The snapback (2021–22): most of the gap closes in eighteen months. Her net result to 2026: roughly world-matching with a wilder path: the honest median outcome of a well-run factor bet.

Her colleague who bought the same fund in mid-2021 (after the feast, on the feast's evidence) and sold in 2023's chop realized the mirror image. Same factor, same fund: the protocol was the return.