Blue vs White Collar

Fifteen BLS-sourced labor-market metrics, split across two cohorts: blue-collar (manufacturing, construction, mining, production wages, less-than-HS unemployment) and white-collar (financial, professional-business, information employment, management + sales earnings, bachelors+ unemployment). For each metric, compute the Democratic-president and Republican-president averages and the gap.

A metric is “won” by the party with the better direction-aware average (higher CAGR for employment/earnings; lower average for unemployment). Ties and insufficient-data metrics are not counted.

Jobs and earnings for manufacturing, construction, mining, and non-supervisory production workers, plus unemployment for workers without a high-school diploma. The winning cell on each row is highlighted.

Jobs and earnings for financial, professional-business, and information-sector employees, management + sales occupations, and unemployment for workers with a bachelor’s degree or higher.

Donald J. Trump II ‡ is 1.79 years into his term (2025-01-20 – today). Current CAGRs across the 15 metrics (direction-aware, see table below). This row updates nightly as fresh BLS releases arrive and will enter the historical D/R averages once the term completes.

The cohort split. Blue-collar metrics measure jobs and wages for workers in goods-producing industries (manufacturing, construction, mining, utilities) plus production + non-supervisory earnings and less-than-HS-degree unemployment. White-collar metrics measure financial, professional-business, and information-sector employment plus management + sales occupational earnings and bachelors+ unemployment. Each metric is explicitly tagged to one cohort in the source data.

Win rule. Within each metric, the party with the better direction-aware average wins. For CAGRs and employment growth, “better” = higher. For unemployment averages, better = lower. Ties and insufficient-data metrics are not counted. The headline tally (D 6 / R 3 on blue; D 3 / R 3 on white) is a simple metric count, not weighted by sample size or statistical significance.

Sample. 28 presidential terms covering roughly 1933 – today. Older terms have smaller n per metric because many BLS series start post-1946 (JOLTS, bachelors+ unemployment) or post-1960 (some industry subsegments). Term-level averaging uses CAGR for level series and simple time-weighted mean for already-ratelike series (unemployment).

Data source. All 15 series pulled from FRED (Federal Reserve Economic Data), stored in presidential_performance.db, regenerated monthly on first-Friday nights with the BLS Employment Situation release and refreshed nightly for timestamp currency.

Caveats.

• Incumbent presidents inherit economies — crediting the party for all employment outcomes conflates policy with cycle.
• Industry composition has shifted structurally (manufacturing share fell from 25% in 1970 to ~8% today). Comparing modern and 1950s manufacturing CAGRs is not apples-to-apples in the usual sense.
• Small-n metrics (n<5 for a party) produce noisy averages that one outlier can flip. Read the n values in parentheses.

• Homeowners vs Renters — same partisan-gap methodology on housing wealth + rent burden.
• Trump vs Biden — Matching Window — T1/B/T2 head-to-head across 85 metrics in 14 cohorts.
• Full-Term Ranking — all-metric rank of every president since Hoover.
• Back to Presidential Performance library