How the score works
The economy score condenses dozens of official indicators into one number from −100 (stressed) to +100 (strong). It's a transparent formula, not a forecast and not an AI judgment: the same inputs always give the same score.
1. Each indicator gets a score
Every scored indicator is rated from −100 to +100 by combining up to two signals:
- Momentum. Where the indicator's recent level (or its change over 6 to 12 months) ranks against its own last 20 years. Ranking each series against its own history puts job gains, mortgage rates and consumer sentiment on the same scale, and adapts as what's “normal” shifts over time.
- Benchmark. Distance from an economic reference point, where one exists: the Fed's 2% inflation target, the Sahm rule's 0.5-point recession trigger, a yield curve below zero, job gains of about 150,000 a month.
Each indicator's settings and the reason it counts are documented in the project's scoring configuration.
2. Indicators roll up into sectors, sectors into the economy
A sector's score is the weighted average of its indicators. The economy score weights the sectors as below. The Federal Reserve is shown as a policy stance (tightening, on hold or easing) rather than scored, because it reacts to the rest of the economy.
| Sector | Weight | Score now |
|---|---|---|
| Inflation | 20% | +3 |
| GDP | 15% | +41 |
| Jobs | 15% | +6 |
| Consumer | 10% | −21 |
| Employment | 10% | +20 |
| Recession risk | 10% | Low |
| Bonds & mortgages | 5% | −89 |
| Energy | 5% | −96 |
| Housing | 5% | −22 |
| Manufacturing | 5% | +36 |
3. Does it track the economy?
Applying the same formula to every month since 1990 and comparing it with official recession dates (NBER):
Every recession month since 1990 scored below zero, and the score was already falling heading into 2001 and 2008. The 2020 recession was a sudden shock that no indicator foresaw. This check runs as an automated test every time the data updates.
Limitations
- The history uses today's revised data, so it shows the score follows the business cycle, not that it would have predicted recessions in real time.
- The weights and benchmarks are judgment calls; they're documented and easy to change.
- It describes conditions now. It is not a forecast, and it is not financial advice.