Two economies.
One number.
Updated weekly.

The Bifurcation Index (BFX) measures how differently wealthy and working Americans are experiencing the same economy - in a single number from 0 to 100.

Built from 33 U.S. government data series. No paid subscriptions. No black boxes. Every Friday after market close, the index updates automatically and the weekly report lands in your inbox.

Subscribe Free Read the Methodology
BFX 57.2
Spending 80.6
Labor 39.4
Housing 50.2
Financial 54.5
Sentiment 61.1
3-mo Δ −7.6
As of Mar 14, 2026
The Reality Behind the Averages

GDP is growing. The unemployment rate is 4.3%. Consumer spending is "resilient."
Here's what those numbers aren't telling you.

In Q2 2025, the top 10% of Americans accounted for 49% of all consumer spending - up from 35% in the early 1990s. The aggregate data looks fine because the top of the distribution is pulling it up. The bottom half isn't in the same economy. The BFX is built to see both.

Top 10% share of spending
49%
Up from 35% in 1990 · Moody's, Q2 2025
Bottom 50% share of wealth
2.5%
Fed Distributional Financial Accounts, Q3 2025
Gini coefficient
60yr
high
Inequality at generational extreme · U.S. Bank, 2026
Top 10% own of equities
93%
Fed DFA · Why bull markets widen the gap

Fed Chair Powell called it the "K-shaped recovery" after COVID. One arm of the K goes up - asset prices, financial markets, luxury spending. The other arm goes down - food insecurity, credit card debt, consumer pessimism. Aggregate indicators average the two arms together and declare everything fine. The BFX doesn't average. It measures the distance between the arms - every week, back to 2008, in a single number calibrated to the pre-pandemic norm.

Current Reading
57.2
−7.6 pts
vs. Dec 2025
Above Baseline

A score of 57.2 means the gap between wealthy and working Americans is moderately above the pre-pandemic historical average. Think of 50 as "normal" - the average divergence measured from 2008 through 2019. Above 50, the gap is wider than that norm. Below 50, it's narrower.

The index dropped 7.6 points over the past three months - down from 64.8 in December. A falling reading is not inherently good news. It means the gap is compressing, but the reason matters. This drop was driven almost entirely by the Sentiment pillar collapsing 29 points as financial conditions tightened and consumer confidence deteriorated following the U.S. military conflict with Iran in late February. The gap narrowed because Wall Street caught down to Main Street - not because lower-income households improved. Compression driven by stress at the top is very different from compression driven by gains at the bottom.

Dec 2025
64.8
3 months ago
All-time high
65.1
Nov 2025
All-time low
33.7
May 2020
Compressed Baseline (50) Extreme
02040506080100
Score bands - constructed backtest
Extreme 85 – 100
High 70 – 85
Above Baseline ← current 55 – 70
Baseline 45 – 55
Below Baseline 30 – 45
Compressed 0 – 30
Five Pillars

What goes into the number

The BFX measures divergence across five domains of economic life. Each pillar is scored 0–100 on the same scale and weighted equally in the composite. 50 = the 2008–2019 pre-pandemic average for that pillar.

Consumer
Spending
80.6
Above Baseline
Nov 2025 +1.7 pts ↑ 79.0

This pillar compares discretionary spending - restaurants, electronics, furniture - against necessity spending: groceries, pharmacies, discount stores. A higher reading means the top of the income distribution is pulling further ahead of the bottom in how much they're spending on non-essentials. At 80.6, that gap is near its widest on record. The Fed's March 2026 Beige Book described it directly: high-end spending held firm while lower-income consumers pulled back and shifted to cheaper alternatives. Upper-income households are carrying the "resilient consumer" narrative almost entirely.

Labor
Market
39.4
Below Baseline
Nov 2025 −2.0 pts ↓ 41.4

This pillar tracks three measures: the wage gap between finance workers and hospitality workers, the unemployment gap between college graduates and high school dropouts, and the ratio of high-wage to low-wage job creation. A higher reading means those gaps are widening. At 39.4, the pillar is below the pre-pandemic norm - the 2021–2023 surge in service-sector wages compressed the pay gap between finance and hospitality. That's a real, accurate signal. What the pillar doesn't capture: the BLS reported that labor's share of GDP hit its lowest point in 78 years in 2025. Workers overall are getting a smaller slice of a larger pie.

Housing
50.2
At Baseline
Nov 2025 −2.8 pts ↓ 53.1

This pillar measures three things: whether luxury homes are appreciating faster than entry-level homes (Case-Shiller tier spread across four major metros), the national homeownership rate, and whether shelter costs are rising faster than wages. A higher reading means housing wealth is concentrating upward - luxury appreciating faster, fewer people able to own, renters falling further behind. At 50.2, the tier spread is back at the historical norm after cooling under sustained rate pressure. The pillar moved down 2.8 points this quarter. The spread narrowing doesn't mean housing is affordable - just that the divergence isn't widening further right now.

Financial
Assets
54.5
Above Baseline
Nov 2025 −7.4 pts ↓ 61.9

This pillar combines four inputs: the NASDAQ vs. its own 10-year trend, revolving credit growth, the personal savings rate, and the credit card delinquency rate. A higher reading means asset values are elevated above trend while lower-income households are leaning on debt. The pillar dropped 7.4 points as markets sold off on tariff and geopolitical uncertainty in January–February - the Fed's top 10% felt that. But credit card delinquencies continued rising even as equities retreated. The bottom of the distribution is still under pressure regardless of what the market does.

Consumer
Sentiment
61.1
Above Baseline
Nov 2025 −29 pts ↓ 90.1

This is the story behind the composite drop. This pillar measures the gap between financial conditions - how easy it is to borrow and invest - and how consumers actually feel about the economy. A higher reading means Wall Street and Main Street are diverging: loose conditions for asset holders, stress on the ground. At 90.1 in November, that gap was near extreme. Then the U.S. military conflict with Iran began February 28 - Michigan surveys after that date showed confidence decline that erased earlier gains - while financial conditions tightened sharply as markets sold off. Both sides of the gap moved together. The divergence compressed by 29 points.

Historical Series

Jan 2008 - Mar 2026 · All pre-March 14, 2026 readings are constructed backtest values

BFX Composite
Baseline (50)

All pre-March 14, 2026 values are constructed backtest figures. Subject to look-ahead bias. Not indicative of future readings.

Historical Context

How to read the history

May 2020
33.7
All-time low. COVID stimulus payments briefly compressed divergence - all households suffered simultaneously, and transfer payments narrowed the gap. The gap being small here doesn't mean times were good.
Nov 2025
65.1
All-time high. Equity markets elevated, upper-income spending surging, policy uncertainty rising - classic K-shape conditions. Asset-owners insulated; wage-earners exposed to cost pressures.
2013–2016 Avg
51.5
The compression period. A tight labor market recovery held bifurcation near baseline. The current reading of 57.2 sits 5.7 points above this period's average - a useful benchmark for "normal."
Mar 14, 2026
57.2
Live inception reading. Spending elevated, Labor compressed, 3-month drop of 7.6 points from near peak. The cross-pillar divergence - not the composite - is where the real story lives.
Empirical Validation

Does it measure reality?

Eight association tests were run on the constructed historical series. Results are reported in full - including null findings and circular relationships. Selective disclosure would be incompatible with the credibility standards this index is designed to meet.

V1 · Stress Episode Discrimination
34.4 COVID Shock Mean
61.8 Tariff Era Mean

During COVID, the government sent direct cash payments to nearly all Americans - that temporarily narrowed the gap between rich and poor. Everyone was hurting at once. But when the economy reopened, markets recovered fast and asset prices soared - benefiting mostly wealthy households - while lower-income workers faced years of slow recovery. The index correctly identified both: compression during the shock, then widening as the recovery played out unevenly.

Descriptive · No circularity · Full episode table in §12.2
V2 · Consumer Sector Leading Relationship
−0.28 r at 3-month lead
·
0.00 r at 12-month lead

When the BFX is elevated, stocks tied to discretionary spending - restaurants, retail, travel - have tended to underperform compared to staples like grocery stores and pharmacies over the following quarter. The logic: a bifurcated economy means lower-income consumers are pulling back first, and they make up the customer base for those categories. The relationship fades completely at 12 months, which is what you'd expect if markets are gradually pricing it in.

Statistically significant · Partial circularity noted in §12.3
V6 · Pillar Dispersion vs. Sector Divergence
+0.32 r contemporaneous
·
+0.25 r at 3-month lead

When the five pillars point in different directions - say, spending is elevated while labor is compressed - different parts of the stock market also start diverging more than usual. Think of it as the index's internal tension showing up externally. This is the cleanest result in the test suite because it doesn't share any inputs with the external variable it's being compared against, so there's no risk of circular logic.

No circularity · Significant through 6-month lead · §12.7
All eight tests documented in full - including null results (HY credit spreads: no relationship at any lag, confirming BFX measures something distinct from credit stress) and a failed regime discrimination t-test (p = 0.57, analytically informative) - in the Methodology Paper §12. All results derived from constructed backtest (2008–2026). Subject to look-ahead bias of unknown magnitude. Live empirical testing commenced March 14, 2026.
Methodology

Built on public data.
No black boxes.

The BFX synthesizes 15 analytical components from the Federal Reserve (FRED), Bureau of Labor Statistics, and Bureau of Economic Analysis into a single comparable score anchored to the 2008–2019 pre-pandemic baseline.

Every FRED series ID, every transformation step, and every normalization parameter is documented publicly. Any researcher with a FRED API key can reproduce the full historical series to floating-point precision.

Three adversarial methodology audits were completed before live publication - institutional review, code-vs-paper verification, and final adversarial review. All critical findings were resolved. The full audit trail is in the methodology paper.

By the numbers
33
FRED data seriesAll publicly available U.S. government data. No proprietary or paid sources required for the core index.
15
Analytical componentsAcross five equally-weighted pillars. Each component is stationary by construction - a statistical requirement for the scoring methodology.
144
Months in baseline windowJanuary 2008 – December 2019. Frozen at construction, never updated. 50 = average of this period, by definition.
3
Adversarial audits completedInstitutional, code-vs-paper, and final adversarial review - before a single number was published. Full audit trail in the methodology paper.
0.97+
Sensitivity analysis correlationThree alternative baseline windows (2006–2019, 2010–2019, 2008–2017) all produce r > 0.97 with the primary series.
The Bifurcation Report

Weekly. Free.
Institutional-grade.

Every Friday after U.S. market close - the current BFX reading, what moved it, and what the data is telling you that the headlines aren't.

Free forever. No spam. Unsubscribe anytime.
IMPORTANT DISCLOSURES: All BFX readings prior to March 14, 2026 represent a constructed backtest. Backtested results are inherently subject to look-ahead bias, data-mining bias, and data vintage effects. The BFX is not an investment product and does not constitute investment advice. No performance claims are made. The 2008–2026 historical series was produced by applying the finalized v5.2 methodology retroactively. © 2026 Bifurcation Analytics. All rights reserved. Trademark registration pending.