Federal funding · Follow-up · Methodology

Federal transit and aviation dollars don't rescue the federal-spending channel for DBE density — except in DC

Published May 26, 2026 · ~6 minute read

TL;DR

Our first /data post tested whether FHWA per-capita apportionment predicts state-level DBE density and found r = +0.06 (p = 0.67). We left open the question of whether adding FTA (transit) and FAA (aviation) funding would rescue the spending channel.
We now have the FY2024 FTA and FAA data. Adding them gives a significant all-states correlation: r = +0.29 (p = 0.04). So at first glance the spending channel looks rescued.
Except DC is doing all the work. DC has the highest combined per-capita federal-aid transportation funding ($967/cap, driven by WMATA-related FTA flows) and the highest DBE density (264/100k). Without DC, the combined correlation collapses to r = +0.05 (p = 0.75)— back to essentially zero, same as FHWA-alone.
FTA alone is the only program with a notable correlation,at r = +0.62 (p < 0.001) all-in, but again DC drives it: ex-DC, FTA-only is r = +0.13 (p = 0.36).
Original conclusion stands: outside the federal-enclave region, per-capita federal-aid transportation funding doesn't predict DBE density. The DC-metro federal- contracting cluster is the spending-channel story, and it's a story about one place, not a national pattern.

In our first /data post we asked whether FHWA per-capita apportionment predicts state-level DBE density. The answer was decisive: r = +0.06 (p = 0.67), and r = +0.01 excluding DC. We documented one obvious limitation in that post's caveats section: FHWA is just one of three federal-aid transportation programs (highway, transit, aviation) that carry DBE goal requirements. The other two — FTA Section 5307/5311/5337/ 5339/5310 and FAA Airport Improvement Program (AIP) grants — could carry the spending signal that FHWA alone missed.

FTA and FAA aren't indexed in USAspending in any useful way (we documented why in the NAICS market-saturation post). The data lives in agency-published apportionment tables that have to be downloaded manually. We did that, parsed the three programs into per-capita figures, and re-ran the test.

Results

FY2024 totals by program (national):

FHWA: $54.61B (Highway Statistics Table FA-4)
FTA: $13.91B (Full-Year Apportionment State Totals across Sections 5303/5304/5307/5310/5311/ 5329/5337/5339/5340)
FAA AIP: $6.65B across the 50 states + DC (Total Amount across all grant types — Entitlement, Discretionary, Supplemental, ATRM, AIG, etc.). FY2024 FAA awarded ~$140M in additional AIP grants to US territories (Puerto Rico, Guam, etc.) that are excluded here for consistency with the FHWA and DBE-side data, which are 50+DC only.
Combined: $75.17B (50+DC)

Pearson r of DBE per 100k vs each program's per-capita apportionment, with and without DC:

StateDC (highlighted as outlier)OLS fit (all 51)OLS fit (ex-DC)

Each panel plots one funding source (or the combined sum) on the x-axis against DBE firms per 100,000 residents on the y-axis. The gray dashed line is the ordinary least-squares fit across all 51 jurisdictions; the rust dashed line is the fit excluding DC. Hover any dot for the state, per-capita funding, and DBE density. Pearson r and p-value shown in the top-left of each panel.

What the chart shows

Top-left panel (FHWA only): the original null finding. r = +0.06 across all 51, drops to +0.01 ex-DC. No relationship.

Top-right panel (FTA only): looks like a strong relationship at first — r = +0.62 (p < 0.001). But notice the single outlier in the top-right: DC at $654 FTA per capita, 264 DBE per 100k. That's WMATA: DC's tri- jurisdictional transit system pulls a large FTA apportionment relative to DC's small population. Without DC, the FTA correlation drops to r = +0.13 (p = 0.36) — modest and not statistically significant. There's a real but weak transit-heavy-state signal (NY, NJ, MA all sit slightly above the OLS line for non-DC), but it isn't the rescue we were looking for.

Bottom-left panel (FAA AIP only): no relationship at all (r = +0.02 all-in, +0.09 ex-DC). FAA AIP per-capita is dominated by small rural states (Alaska at $438/cap, North Dakota $99, Vermont $94, Hawaii $91, Wyoming $86) for the same floor-protection reasons FHWA per-capita is. None of those states have high DBE density.

Bottom-right panel (Combined FHWA + FTA + FAA): the headline. All-in r = +0.29 (p = 0.04), statistically significant. Ex-DC r = +0.05 (p = 0.75), back to no relationship. The combined-spending channel "rescue" we hypothesized is real if you count DC, and disappears if you don't.

What this means

Two readings, depending on how you treat DC:

DC included: federal-aid transportation funding per capita does predict DBE density across the full set of 51 jurisdictions, primarily through the FTA channel. The relationship exists. Federal transit dollars flow to DC (via WMATA), and DC has high DBE density. The spending-channel hypothesis is rescued, with FTA as the mechanism.
DC excluded: the relationship vanishes. Across the 50 states, per-capita federal-aid transportation funding does not predict DBE density. The all-in correlation was driven by one data point.

Which reading is "right" depends on what you're trying to learn. If the question is "does federal-aid transportation funding shape DBE density?" the honest answer is "yes, but only via the DC-metro federal- contracting cluster, and the relationship is real for that cluster." If the question is "does it predict DBE density across the country?", the answer is no.

We covered the WMATA-as-DBE-hub story in our regional transit authorities post: 83.2% of WMATA-certified DBEs are headquartered outside DC. WMATA's FTA apportionment supports a contracting ecosystem that spans DC + Maryland + Virginia, and the DBE registrations follow that footprint. This new analysis confirms the same pattern from the funding side: when DC looks "rich" in federal transportation funding per capita, it's the regional transit authority inflating the per-capita figure, and that authority is exactly what creates the contracting environment in which DBE certification gets pursued.

Where the FTA-only signal is interesting

Even excluding DC, FTA per capita shows a weak positive trend with DBE density (r = +0.13). Look at the top FTA- per-capita states ex-DC:

New York: $116 FTA/cap, 21 DBE/100k
Alaska: $113 FTA/cap, 86 DBE/100k (small-state-formula outlier)
New Jersey: $90 FTA/cap, 23 DBE/100k
Massachusetts: $78 FTA/cap, 20 DBE/100k

The transit-heavy northeastern corridor states sit above average on both axes. Not enough to be statistically significant at n=50, but enough to register as a directional pattern. If we ever get state-level data for FY22 and FY23 FTA apportionment (the agency doesn't publish a per- state rollup for those years), a multi-year analysis might sharpen this.

Methodology

FHWA: FY2024 Highway Statistics Table FA-4, APPORTIONED TOTAL column (used directly, since for FY2024 the column equals the sum of program columns). National total: $54.61B.
FTA: FY2024 Full-Year Apportionments State Totals, downloaded from transit.dot.gov/funding/apportionments. The State Totals workbook aggregates across all major formula sections (5303, 5304, 5307, 5310, 5311, 5329, 5337, 5339, 5340). National total: $13.91B.
FAA AIP: FY2024 Airport Improvement Program grant detail report from faa.gov/airports/aip/grant_histories. Airport-level "Total Amount" column summed to state. Includes Entitlement, Discretionary, Supplemental Discretionary, ATRM, AIG, FCT, and other AIP program components for FY2024. Total across the 50 states + DC: $6.65B. The full FAA file (including ~$140M in territory grants) totals $6.79B; territories are excluded here for consistency with FHWA and the DBE source data.
Population: US Census Bureau Population Estimates Program (PEP) vintage 2024 state-level totals. Same as the first post.
DBE counts: DBE Source database, May 2026 snapshot. Companies tagged by HQ state in companies.state.

Frequently asked questions

Does federal transportation funding predict DBE density?

Including DC, yes — combined FHWA + FTA + FAA per capita correlates with DBE density at r = +0.29 (p = 0.04). Excluding DC, no — the correlation collapses to r = +0.05 (p = 0.75). The relationship is a DC-metro story, not a national one.

Which federal program correlates most with DBE density?

FTA (transit), at r = +0.62 all-in — but this is driven primarily by DC's very high per-capita FTA apportionment (WMATA-related). Excluding DC, FTA per capita correlates with DBE density at r = +0.13 (p = 0.36, not significant). The transit-heavy northeast corridor (NY, NJ, MA) sits modestly above the OLS line without DC, suggesting a weak but real transit-density signal beyond just the federal enclave.

Why does DC have such high FTA funding per capita?

DC's FY2024 FTA apportionment is $459M against a population of 702K — $654 per capita. This reflects the Washington Metropolitan Area Transit Authority (WMATA), which is administered through DC's FTA allocation but serves a tri-jurisdictional area (DC + Maryland + Virginia). DC's small population inflates the per-capita figure dramatically.

What about FAA per capita? Alaska is at $438 — should that show up?

Alaska's extreme FAA per capita reflects the state's heavy reliance on aviation for inter- community transportation, and the federal AIP program supports a large number of small Alaska airports. Despite this, Alaska's DBE density is only 86/100k — lower than DC (264) or MD (231). High federal transportation funding per capita doesn't automatically produce high DBE certification rates. Alaska's contracting model and certification program structure differ from the DC-metro region.

Caveats

We used FY2024-only for FTA and FAA. The first post's addendum tested whether FHWA 3-year averaging changed anything (it didn't — r identical to three decimal places), and FTA + FAA apportionment formulas are similarly stable across years, so single-year usage is unlikely to be material. FTA doesn't publish a per- state rollup for earlier fiscal years, so multi-year averaging would require manual UZA-to-state crosswalks that we judged not worth the work given the FHWA averaging evidence.

FAA AIP "Total Amount" includes some grant streams that aren't strictly DBE-applicable (e.g., supplemental COVID-era streams). Net effect on per-capita rankings is small — the floor-protected small states still dominate FAA per-capita whether or not we include those components.

n = 51 is statistically thin. The DC-as-outlier finding isn't a methodological artifact — it's a real feature of how the data distributes — but we wouldn't over-interpret any p-value from a sample this small.

Data snapshot: May 26, 2026. Companion analyses: DBE density by state (original FHWA-only finding), regional transit authorities (WMATA story), county-level density.

← Back to Data