← All strategies

Growth-trend timing (unemployment)

▲ Promising

Index timing — holds SPY or T-bills

The trend filter, but it only steps aside when the economy is weakening too — unemployment rising above its 12-month average. Market dips without a recession behind them are ridden out.

View 6-month backtest → View 1-year backtest →

New backtests are run by the owner.

How it works

An illustration of the rules on made-up prices, not real data.

Long-run result — 1994–2026, SPY with T-bill cash

Annual return

12.0%

SPY 10.9%

Worst drawdown

-33.7%

SPY -55.2%

Beat SPY with ~40% less drawdown by stepping aside in 2001–02 and 2008; slightly behind since 2010.

Growth of $100,000 vs the same amount in SPY, 1994–2026

From the latest backtest run (to 2026-10-19): 11.9%/yr vs SPY 10.8%/yr; worst drop -33.7% vs -55.2%. Log scale, so equal percentage moves look the same size. Hover for values.

Rules

  1. At each month-end: is SPY below its 10-month average?
  2. And is the latest published unemployment rate above its 12-month average? (FRED UNRATE)
  3. Both → sell and hold T-bills; otherwise hold SPY

Assessment

  • + Targets the trend filter's weakness — whipsaws in markets with no recession (2022)
  • + Unemployment trend has flagged every US recession since 1945
  • + Few trades; months-to-years horizon
  • − False recession alarms still happen
  • − Unemployment turns late, so the first leg of a crash is often taken
  • − Rests on a handful of recessions — small sample
Last 6-month backtest: 07 Oct 2026 — SPY 15.8% Last 1-year backtest: 07 Oct 2026 — SPY 17.7%