Key Takeaways
Powered by lumidawealth.com
- The economy neither boomed nor collapsed: growth stayed resilient, helped by AI investment and strong asset markets.
- Tariffs raised meaningful revenue, but nowhere near enough to replace income taxes; legal risk could force refunds and cut collections.
- Inflation rose in tariff-exposed goods, but broad inflation stayed around ~3% as housing/energy dynamics and policy uncertainty muted pass-through.
- Manufacturing hasn’t revived: activity stayed in contraction, and volatile tariff policy discouraged long-horizon capex and reshoring.
- The trade deficit didn’t “flip”: it swung with pre-tariff stocking and subsequent normalization; structural savings/investment forces still dominate.
- Labor is softening: job growth slowed and unemployment rose, with manufacturing jobs down on net, though attribution is messy.
What happened?
After April’s tariff escalation, forecasts split between recession warnings and a manufacturing renaissance. Neither fully materialized. GDP stayed strong (notably 2Q 2025), but reshoring and factory momentum remained weak. Tariff revenue surged, while inflation effects were concentrated in specific imported goods categories rather than the overall CPI.
Why forecasters missed it
- AI capex became a larger growth driver than expected, offsetting tariff drag.
- Tariff rates moved around (delays, rollbacks, legal uncertainty), changing behavior and slowing price pass-through.
- Import substitution (shifting sourcing and product mix) lowered the “effective” tariff burden versus headline rates.
- Macro inflation was driven more by housing/energy than tariffed goods.
What to watch next
- Supreme Court ruling on tariff authority: revenue path, refund risk, and whether tariffs shift to other legal tools.
- When inventories normalize and contracts reset: delayed inflation pass-through could still surface.
- Whether policy stability improves enough for manufacturers to commit to multi-year reshoring and capacity builds.
- Labor-market revisions and data quality: late survey responses and model assumptions could change the picture.
