Plain numbers, plain assumptions. The whole point of an index is that you can argue with it. Here's how the score gets built, every input it uses, and the things we know it doesn't capture.
We take five line items at the metro level, normalize each one to a 0–1 sub-score over a sensible range, weight them, and sum to a 0–100 Quiet-Broke score. Higher = more squeezed.
score = 100 × (0.40 × housingn + 0.25 × taxn + 0.15 × childcaren + 0.10 × healthcaren + 0.10 × transportn)
Each xn is the line item as a share of gross income, clamped to a normalization range tuned to where $400K households actually sit. Housing, for example, normalizes between 6% of gross (a Cleveland-style baseline) and 20% of gross (a coastal-metro ceiling). Above 20% you're saturated at 1.0 — additional rent doesn't make you "more than maximally squeezed."
The weights aren't from a regression — they're a deliberate editorial choice. Housing gets 40% because it's the line that moves the most between metros and the one HENRYs report feeling most. Tax gets 25% because it sets the post-tax denominator everything else fits inside. Childcare gets 15% because at the modal household it's the third-largest dollar line. Healthcare and transport split the remaining 20% because they're meaningful but relatively flat across metros.
If you re-ran the index with equal 20% weights, the top of the list would still be New York / San Francisco / San Jose / Boston, and the bottom would still be Cleveland / Houston / Nashville. The weighting changes the middle, not the shape.
All inputs and per-metro numbers live in docs/data/cities.json in the source repo. The numbers refresh monthly.
This is the part most data projects skip. It shouldn't be.
Wildly variable, household-specific, and load-bearing for a meaningful share of HENRY households in NYC, SF, LA, and DC. We assume public school. If you're paying $35K–$60K per kid per year, your real-life Quiet-Broke score is materially worse than this calculator shows.
Same problem. These are catastrophic-tail line items that can dominate a household budget for a year or a decade. We can't fairly average them into a metro number.
We use 3BR rent as the housing input because it's apples-to-apples across metros. A long-time owner with a sub-4% mortgage is paying nothing close to today's market rent — they're functionally less squeezed than our index suggests. A 2024-buyer with PITI on today's prices is more squeezed.
Two HENRYs at the same gross can spend ±$60K differently on discretionary. The index is about the floor of the bill, not the ceiling. If your discretionary is eating you alive, that's a real problem — it's just not the problem this number is trying to measure.
We treat $400K as cash-equivalent. In real life a lot of HENRY gross is unvested or single-stock RSU, which has its own risk and tax timing dynamics we're not modeling here.
An honest list of where this v0.1 could mislead you:
This is a one-page index from one publication, version 0.1, refreshed monthly. It's a starting point for a conversation, not the IRS. Henry Finance writes about money decisions for HENRY households; this is a research project that fell out of one of those decisions and turned out to be more interesting than the original article.
If you think a number is off — and you have a source — email us and we'll fix it publicly. Material corrections get a visible note on the page with the date.
One free issue every Tuesday and Friday. Frameworks, not formulas. We send the next deep-dive on the highest-leverage line items on this index — childcare arbitrage, the SALT/AMT shift, and house-rich/cash-poor traps.