I spent fifteen years running retail assets on this corridor for institutional owners at JLL. In that time I watched a lot of very good site-selection work produce the wrong answer here, and it was almost never because the team was careless. It was because the standard toolkit encodes assumptions that are true almost everywhere in the United States and false in a border county.
This page is the list of those assumptions, with the specific mechanism for each and what to check instead. It is not a criticism of anyone's model. It is a list of the places where this market breaks one.
1. A drive-time ring treats an international boundary as ordinary geography
This is the largest single error and it runs in both directions at once.
A ten-minute ring drawn from a site on South Bibb includes Piedras Negras. Standard demographic packages will either populate that area with nothing — because U.S. census geography stops at the river — or populate it with Mexican data that has no household income attached. Either way the model is wrong: the first understates the market badly, the second overstates the certainty.
And the boundary is not a normal edge. Crossing it involves documents, an inspection and a wait. Two sites four miles apart, one near a bridge and one on the loop, capture completely different flows — and no drive-time engine models a port of entry.
What to check instead: the measured crossing data. About 6.96 million inbound person-crossings in 2025, of which 1.31 million on foot, up 52.5% since 2019, while personal vehicles remain below 2019. That split is the trade area, and it is public. See the crossing data in full.
2. Income understates spending — by roughly half
Maverick County per-capita income is $21,211 against $40,752 for Texas. A model that sizes demand off local income will size this market at roughly half.
The county nevertheless rings up essentially exactly the Texas average retail sales per resident — $6,995 against $7,073. On an income-adjusted basis that is about 1.9× more retail volume than local buying power supports.
The caution that keeps this honest: on the same basis Uvalde County scores 1.74 and it is not a border market. County seats import demand from their hinterland everywhere. Do not attribute the entire gap to Mexico — the data cannot support that, and this dataset cannot separate cross-border shoppers from commuters, tourists, pass-through traffic and business purchases.
3. The city-level pull factor flatters the market and it is the figure most often quoted
At city level Eagle Pass captures 1.90× the Texas average per resident. At county level it captures 0.99×.
The difference is not a measurement dispute. Eagle Pass city holds 94.4% of the county's retail sales but only 49% of its population, so a city-level ratio counts county residents living outside the city limits as imported demand. Both numbers are real. Only one of them is a trade area.
A deck that shows 1.90 without 0.99 beside it is showing one of two real numbers. Ask for both.
4. The market stopped growing and the headline revenue figure hides it
City taxable sales grew 25.9%, 14.0%, 7.9% and 3.7% in 2021 through 2024 — then fell 0.5% in 2025. Retail trade specifically fell 2.9%. In nominal dollars, a flat year is a real decline of roughly three percent.
Meanwhile the City's own sales tax allocation payments set a record in 2025. Both are true. The allocation payments imply about 38% more taxable sales than the outlet-level report shows, and that gap has widened since 2019 — most plausibly destination-sourced e-commerce and remote-seller tax flowing to Eagle Pass addresses. That is revenue to the City. It is not sales through a store.
What to check instead: the outlet-level taxable sales series, not the allocation series, whenever you are underwriting a building.
5. Texas is a non-disclosure state, so there is no comp set
No sale price is public in Texas. On top of that, in this market:
No occupancy figure is publicly available for any Eagle Pass retail property, including the enclosed mall. No gross leasable area, year built or occupancy is published for the market's principal non-mall multi-tenant center. Asking rents are not published. The one recent institutional trade — the mall at a reported $10.04 million in August 2024 — is known only through partly paywalled trade coverage, and the implied per-foot price does not reconcile with the owner's own square footage claim.
And the county's public parcel data is a frozen extract: every record carries a 2022 appraisal-year label, but no deed in the file postdates September 2021. Anyone citing it as current ownership is citing a five-year-old snapshot.
What to check instead: construction registrations. They are free, dated, name the owner and the tenant, and state the square footage and the cost. They are the closest thing to a transaction record this market has.
6. A low category concentration is not automatically an opportunity
Health and personal care runs at 0.49 of the national employment concentration here. Sporting goods, hobby and books also 0.49. A gap-analysis model reads those as unserved demand and recommends a store.
Maybe. Or the category is being served across the river, or online. On this corridor both are live explanations, and no dataset distinguishes them. The Dallas Fed's own explanation for the long decline in border retail capture is the expansion of big-box retail inside Mexico — Walmart went from 204 Mexican stores in 1998 to 3,154 by 2024.
What to check instead: whether the category exists in Piedras Negras. That is a drive, not a data pull, and it is the single highest-value hour in the whole exercise.
What the pipeline tells you that the model will not
Across 2023 to August 2026, every retail construction registration in Maverick County is a pad, a discount box, a car wash, a quick-service restaurant or a truck stop. There is exactly one speculative multi-tenant retail building in the entire record — a 10,740 square foot shell on El Indio Highway, still at plan review.
Read that as the market's own revealed judgment. Capital that knows this market is building single-tenant net-lease product and nothing else. That is either the opportunity or the warning, depending on what you are proposing — but it should be an input either way.
The one genuine outlier is Olive Garden, registered December 2025 at 2071 North Veterans, $3,274,500 and 7,828 square feet, scheduled April to October 2026. It would be the first full-service national casual-dining chain of its kind here. It is authorized and designed, not confirmed built.
The short version
The reason an out-of-town model misses here is not that the data is bad. It is that the three things that decide this market are all invisible to it — an international boundary that a drive-time ring cannot see, imported demand that local income cannot explain, and a comp set that a non-disclosure state does not produce.
Everything on this page is public and free. What it cannot tell you is which of two corners captures the walk-up flow, whether the category you are missing exists in Piedras Negras, and what a landlord here will actually accept. Those are the questions worth paying for, and they are answered on the ground.
Common questions
Why do retail site-selection models get Eagle Pass wrong?
Six recurring mechanisms: drive-time rings treat an international boundary as ordinary geography; local income understates actual spending by roughly half; the widely quoted city-level pull factor of 1.90 falls to 0.99 at county level; the market stopped growing in 2025 while municipal sales tax receipts hit a record; Texas non-disclosure means there is no comp set and no published occupancy for any property here; and a low category concentration may mean the category is served across the river rather than unserved.
What is the correct trade area for an Eagle Pass retail site?
There is no single ring. The county holds about 58,800 people and the four nearest counties about 79,000. Across the river, nine adjacent Coahuila municipios hold about 272,600, of which Piedras Negras is 176,327. The functional measure is crossing data — about 6.96 million inbound person-crossings in 2025 — and the split between pedestrian and vehicle flows matters more than the totals.
Are there retail comparables available in Eagle Pass?
No. Texas does not disclose sale prices, no occupancy figure is published for any Eagle Pass retail property including the mall, and asking rents are not published. Construction registrations — which are free, dated and state owner, tenant, square footage and cost — are the closest available substitute.
Is anyone building multi-tenant retail in Eagle Pass?
Effectively no. From 2023 through August 2026 the registered pipeline is pads, discount boxes, car washes, quick-service restaurants and one truck stop, plus a single 10,740 sf speculative shell on El Indio Highway that remains at plan review.
Sources. Texas Comptroller of Public Accounts, Quarterly Sales Tax Historical Data (7z4d-yf2c); U.S. Census Bureau ACS 2020–2024 5-Year and Population Estimates Vintage 2025; U.S. Bureau of Labor Statistics QCEW area 48323; U.S. DOT Bureau of Transportation Statistics Border Crossing/Entry Data (keg4-3bc2); Texas Department of Licensing and Regulation TABS registrations; Federal Reserve Bank of Dallas, Southwest Economy (January 2026). Verified as of August 2026; project statuses change and every figure should be re-verified before it is relied upon.