# Responsible America — Federal Contracts Tracking Methodology

Status: architecture document plus a first real, hand-verified seed
dataset (added 2026-08-11). Companion to
`docs/public-scorecard-methodology.md` and
`docs/political-money-tracking-methodology.md` (same "architecture before
data" approach, and the same real-data-civic-tools batch).

Last reviewed: 2026-08-11.

## 1. What this is, and what it is not

This is the data architecture for tracking federal contract awards, built
directly on USAspending.gov's own documented data structure (which itself
derives from the Federal Procurement Data System, FPDS) rather than a new
taxonomy invented for this platform. `federal-contracts-tracking.html`
presents the fields that matter and why.

**It is still not a live tracker.** This static site has no USAspending
API ingestion pipeline, no server, and no scheduled data refresh. As of
2026-08-11 it does display real dollar figures for 4 specific named
awards (§2 below) — but every one is a manually researched, individually
sourced, one-time snapshot, labeled with its own "last verified" date.
Nothing on this page updates itself.

## 2. The seed dataset: four named awards

Added 2026-08-11 as this tool's first conversion from architecture-only to
real, source-backed content. Selection criteria and full research notes
are in `research/tool-federal-contracts-seed/research-plan.md`. In short:

- **Lockheed Martin — F-35 Lot 18 production, $11.8 billion (2024-12-20).**
  A sole-source award: the Pentagon's own director of defense pricing
  described F-35 production contracting as "one of the Pentagon's
  largest-ever sole-source contracts." The record's `notes` field explains
  why: Lockheed remains the only qualified F-35 producer following the
  original, competitively awarded 2001 development contract — this
  specific award wasn't competed, but the underlying program was, once.
- **Lockheed Martin — F-35 Lot 18/19 modification, $12.5 billion
  (2025-09-29).** A real modification of the award above, linked via
  `parent_award_id` — the concrete demonstration of this dataset's
  modification-history field. The modification definitizes Lot 18 and
  adds Lot 19 scope; its dollar figure is not new spending disconnected
  from the base award.
- **Boeing — KC-46A tanker development contract, $3.5 billion
  (2011-02-24).** A historical award, clearly dated and never implied
  current, included specifically for genuine `competition_status`
  contrast: this contract was competitively awarded (`full_and_open_
  competition`) after Boeing's successful protest of the Air Force's
  original 2008 award to Airbus/EADS.
- **Leidos — Social Security Administration IT delivery order, ≈$1.06
  billion.** A civilian-agency example showing the `delivery_order`/IDV
  distinction: this figure is one task obligated under a larger vehicle
  (parent IDV `SS001760016`), not that vehicle's full ceiling.
  `competition_status` is left `null` on this record — no source
  confirmed a specific value this pass, and this platform does not guess
  one rather than leave a field honestly unconfirmed.

Every record carries a `fact_vs_inference` sentence: the fact asserted is
always narrow (an award/modification record exists, as reported by an
agency or trade press, as of a stated date). This platform does not infer
waste, favoritism, or misconduct from an award's size or recipient, and
says so on every record.

**Verified seed dataset — last verified 2026-08-11.** This is a
demonstration dataset (4 awards), not a comprehensive one — it covers a
tiny fraction of federal contracting activity and is not refreshed on any
schedule.

## 3. The fields, reusing USAspending's own structure

| Field | Source of the model | Notes |
|---|---|---|
| Recipient | USAspending `recipient_name` / `recipient_uei` | UEI (Unique Entity Identifier) is the current SAM.gov successor to the legacy DUNS number. |
| Agency | USAspending `awarding_agency` | The federal agency that made the award. |
| Amount | USAspending obligation fields | Dollars obligated, not necessarily the full lifetime contract ceiling. |
| Award type | USAspending's own contract-type categories | Definitive contract, purchase order, delivery order, BPA call, or IDV (Indefinite Delivery Vehicle). |
| Period | USAspending period-of-performance fields | Start/end dates. |
| Competition status | GAO's own methodology (see §4) | Not a USAspending field name directly — GAO's competition-rate framework is the authoritative approach reused here. |
| Modifications | USAspending modification-number field | Present when a record represents a change to a prior award, not the original action. |
| Source links | This platform's own evidence system | Every record must cite a real source_id — never asserted without one. |

## 4. Competition status: GAO's methodology, reused

GAO defines a contract competition rate as the dollar amount obligated
annually on competitive contracts and orders, as a percentage of the
dollar amount obligated on all contracts and orders. In one review
(GAO-14-395), GAO found the Department of Defense's competition rate
declined from 62% (FY2009) to 57% (FY2013), varying significantly by
component (Army 66% vs. Missile Defense Agency 29% in FY2013), with "only
one responsible source" as the primary stated basis for non-competitive
exceptions (69% of exceptions in FY2014). This platform's schema
(`data/schemas/federal-contract-award.schema.json`) reuses GAO's own
competition-status categories rather than inventing new ones. Those FY2013
figures are shown as a historical example of the methodology, not as a
current measurement of any agency's present-day competition rate.

## 5. What would need to be true before this becomes a live tracker

A USAspending.gov API or bulk-download ingestion pipeline, a defined
refresh cadence, and a decision about which agencies/award-size thresholds
to track (the federal government awards millions of contract actions
annually — a full mirror is neither feasible nor useful; a scoped,
publicly-documented selection criterion would be needed). §2's four seed
awards already carry a real `source_id` pointing to their own agency
announcement or USAspending.gov award page — the same evidentiary standard
every other record on this platform already carries. What's still missing
is scale and a refresh mechanism, not the evidentiary pattern itself.