The Corridor Price Ladder: Kyrgyzstan-Bound Cars Are Declared at 2.5× Kazakhstan's Value
Georgia's H1 2026 trade tables show a 2.5× spread in declared value per exported car: about $38,600 toward Kyrgyzstan and $15,600 toward Kazakhstan. Per-unit declared value is an aggregate, not a price index, and the tables do not identify the models, ages, condition or final destination behind the difference.
Five corridors, one surprising column
Georgia's recorded passenger-car exports in the first half of 2026 went overwhelmingly to five destinations: Kazakhstan (13,068 cars), Kyrgyzstan (10,701), Iran (5,086), Tajikistan (2,966) and Azerbaijan (2,456). Our 20 July report covered the volumes and the Iran surge; our 26 July report covered the flow ratio. This piece looks at the column both skipped: declared value divided by units.
That division produces a ladder. Kyrgyzstan-bound cars were declared at about $38,600 each and Azerbaijan-bound at $33,700 — while Kazakhstan-bound cars averaged $15,600, Tajikistan-bound $16,800 and Iran-bound $17,600.
Declared value per exported car by destination, H1 2026
US dollars; declared value divided by recorded units (Geostat external-trade tables). Declared customs values, not transaction prices.
A per-unit value is still not a price index
Our earlier reports warned that average declared value must not be read as a market price, and that warning stands: customs declarations are not sale prices, the mix of model years shifts month to month, and a handful of very expensive vehicles can move a small corridor's average.
The measured difference is 2.5× between two large destination aggregates. The table establishes that their average declared values differ; without vehicle-level values or model mix, it cannot show how much comes from composition, a skewed distribution or other declaration effects.
A per-unit value is still not a price index
Our earlier reports warned that average declared value must not be read as a market price, and that warning stands: customs declarations are not sale prices, the mix of model years shifts month to month, and a handful of very expensive vehicles can move a small corridor's average.
The measured difference is 2.5× between two large destination aggregates. The table establishes that their average declared values differ; without vehicle-level values or model mix, it cannot show how much comes from composition, a skewed distribution or other declaration effects.
Declared value per exported car by destination, H1 2026
US dollars; declared value divided by recorded units (Geostat external-trade tables). Declared customs values, not transaction prices.
What the ladder is made of — and what we cannot see
Trade tables stop at units and dollars; they do not name models, ages, conditions or final buyers. The vehicle composition behind the ladder is therefore not observable in this dataset, and labels such as newer, mass-market or premium would require a separate vehicle-level source.
The defensible comparison is limited to destination-level averages: the five recorded corridors have materially different declared value per unit. It does not identify a typical car within any corridor.
The view from the listing floor
The median asking price of an active uncleared AutoBridge listing is $14,500, numerically close to the Kazakhstan and Tajikistan per-unit declared values. This is a scale comparison only: a listing median and a customs average use different populations and denominators.
The data cannot establish whether any corridor is supplied through public listings, direct orders or another channel. Linking the two datasets would require vehicle-level export records.
What this analysis cannot answer
Recorded exports are re-exports in this series; a car recorded toward one country can be resold onward, and declared values can differ from amounts actually paid. Nothing here measures compliance with any country's import rules or sanctions regimes, and we make no such claims. The ladder describes what Georgia's trade statistics record — no more.
Trade figures are Geostat external-trade tables for January–June 2026 as reported by Georgian business media (checked against the same tables used in our 20 and 26 July reports); per-unit values are declared value divided by recorded units, in current US dollars. Declared customs values are not transaction prices. The listing-floor comparison uses active AutoBridge canonical listings, snapshot 3 August 2026; asking prices, not transactions.
Methodology
Geostat external-trade tables for January–June 2026 (passenger-car exports by partner country, units and declared value), as reported by Georgian business media; AutoBridge listing database under the canonical analytics filter for the listing-floor comparison.
34,277 exported cars across the five largest recorded destinations in H1 2026; per-unit values computed as declared value divided by recorded units per destination.
January–June 2026 trade data; active-listing snapshot 3 August 2026.