About Research Education
Build
Millionaire MindsetIn-school programming · 100k+ taught Barrier BreakersPowered by Valmark · $20k+ in prizes Ambassador ProgramPowered by young adults globally NavTechNavigation tech across food, housing, health & jobs
Measure
AI Economic SignalAI exposure meets household stress · monthly Financial Health BarometerReal-time economic pulse · 50 states GitHub ↗Research to reality ResourcesGuides, tools & worksheets
Donate
Blog LinkedIn Contact
AI & The Economic Transition
Building Now

Financial stress shows up before the numbers do.

We're building the signal that catches it — early enough that we can still shape what comes next. FinMango is a nonprofit building financial-health infrastructure for young adults and the communities the economy leaves behind.

See what we're building
The Problem

Most measurement tells you what already happened.

01 / Built for a slower era The official numbers move late. The systems we use to read the economy were built for a different pace. By the time the data confirms a downturn, the stress has been real for months.
02 / Rankings, not reality Exposure isn't impact. Task-automation indices tell you what could be automated. They don't tell you what people can actually access, or who is already feeling the squeeze.
03 / Wages miss the point We measure lived stress. Rent, groceries, bills. The financial pressure households actually carry, not just what shows up in a paycheck.

Two arms. One system.

Measurement on one side, intervention on the other. Both independent, both open. Read the shift early, so the people in its path can get ahead of it — not just ride it out.

Now Building Arm 01 / Measurement The Signal An open, monthly, state-by-occupation read that ties AI exposure to the household financial stress people show online. High-frequency, sub-national, fully reproducible. We're early, and building it in the open. Early build · Open dataset Arm 02 / Intervention The Rails Our diagnostic and routing tools. They turn the signal into help people can actually act on, the moment the stress shows up, matching families to local aid in minutes. Live today · NavTech
See the stress early route people to help
The Research

The plan. Two papers.

This is a proposal, not a result. Nothing here is written yet. We're sharing the plan early to find the collaborators, co-authors, and funders who want to build it with us.

The Discipline

Neither paper claims AI caused anyone's financial distress. This is measurement, not causation. That line is the whole point, and we don't cross it.

01
Paper 1: The open dataset. Descriptive · Reproducible · Where we'd start

An open, monthly, state-by-occupation measure of AI-exposed financial stress, shipped with the pipeline that builds it. The job is to report what the data shows, descriptively. Nothing predictive, nothing causal.

  • Where the stress isWhich occupations and states carry the most AI-exposed financial stress, month to month.
  • Real eventsHow the signal behaved around actual layoffs and restructurings, written up as case studies.
  • The geographyWhere high-exposure stress clusters, and where it doesn't.
  • Validation in the openThe divergence-from-baseline method on display, so reviewers can watch search saturation get handled.
02
Paper 2: The leading indicator. Pre-registered · Falsifiable · Either result is a paper

Does the combined exposure-and-stress signal move before realized occupational unemployment does? We lock the test before we touch the outcomes.

  • The claimThe signal at month t leads unemployment at t+1 to t+3 among high-exposure occupations across states.
  • The testA panel regression, with the method fixed in advance. No quietly picking the lag that looks best.
  • Pre-registration firstIt goes up before we pull a single outcome. That is what makes the result believable.
  • Both ways countPositive, and it's an early warning months ahead of the BLS. Null, and it's an honest bound on what search can tell us.
03
The gap we're filling. The combination is what's new

The ingredients all exist. Nobody has stitched them together.

  • Measuring AI exposureA crowded field. Felten, Eloundou, the ILO, Anthropic's Economic Index, MIT's Iceberg Index. We just use these as an input.
  • Linking exposure to unemploymentHeating up. But most of it measures the present or looks back, and the closest work stops before ChatGPT existed.
  • Reading distress from searchOld news. Choi and Varian, the FEARS index. We borrow the method, with real Health Trends probabilities.
  • The openingA high-frequency household-stress signal, joined to AI exposure, tested as a leading indicator, state-by-month, in the generative-AI era. That's the contribution.
After The Papers

The papers are the foundation. Not the finish line.

The goal was never just to measure the shift. It's to see it early enough to help shape it — to give people, and the institutions around them, the lead time to act on the change instead of just absorb it.

01 / Stand it up A live early-warning system. If the leading-indicator test holds, the signal becomes a public, monthly read that flags labor stress months before the official numbers confirm it.
02 / Close the loop Wire the signal to the rails. When the read spots stress in a place or an occupation, our diagnostic and routing tools point those households to real, local help. Measurement, meet intervention.
03 / Scale it Repeatable by design. The pipeline is open and reproducible. With the right partners, the same read can extend to new regions and new questions, wherever the measurement gap is widest.
Why FinMango

We've done this before.

01
A decade in Ten years building financial-health infrastructure for young adults and vulnerable communities.
02
Global by design A Global Ambassador Program across 12 countries. This isn't just a US story, and we can keep it relevant everywhere.
03
Real scale Programs that have reached 100,000+ students.
04
Rare data Access to actual Google Health Trends probabilities, not the public 0-to-100 index. Hard to replicate.
Get In Touch

Let's build it together.

We're early, and that's the point. If you're funding the response to the AI economic transition, building the infrastructure to measure it, or working to reach the people it hits first, let's talk. We're looking for collaborators and co-authors, not an audience.

Researchers. Funders. Frontline orgs. Co-authors.

Email FinMango Follow the research