Hermès Resale Report — October 2026

Method. Join listing_id across two dates, calculate 100 × (final price / initial price − 1), and take the median of those individual changes; the daily basket uses all registered listings on each date.

Caveat. Changes in USD, without original price history or historical FX. The logger defaults to Birkin and does not store style; a mislabeled historical Kelly exists. Matching selects persistent listings, mixes resellers, and does not measure realized profitability. Coverage, additions, removals, and corrections alter the basket; the recorded number is not the market stock.

Method. Medians and quartiles by model+size across all active priced bags, including special variants; additional Togo contrast within the standard subset.

Caveat. The overall curve mixes leather, condition, color, year, hardware, and variants. Togo groups exclude special variants/HSS, but also do not adjust for remaining characteristics. Asking price does not equal realized price or a causal size premium.

Method. Select the five most numerous model+size+leather+hardware combinations; interpolated p10 and p90 per listing and spread 100×(p90/p10−1). Reseller median comparison restricted to n≥15 per reseller within the spec.

Caveat. Matching spec does not mean an identical bag: color, condition, year, provenance, and services vary. Hardware is also inferred from the title. Dispersion reflects listings across resellers, without authentication or physical deduplication across sites.

Method. Within the standard subset, classify crocodile, alligator, ostrich, and lizard as exotic according to the normalized leather field; compare medians by model+size against identified regular leather.

Caveat. The mix of species, color, condition, year, and reseller varies. No adjustment is made for rarity or quality; the descriptive difference in asking price does not measure return or causal effect of the material. Unknown material is excluded.

Method. Cohort of retained bag documents with a valid price and created_at between 2026-08-09T06:18:21.107277+00:00 and 2026-09-02T11:31:08.261615+00:00; count stale_reason=source_missing_two_successful_runs and stale_at−created_at≤30 days, over all retained eligible additions, whether active or retired.

Caveat. Retirement ≠ sale: it may indicate delisting or absence from the feed. The denominator does not reconstruct the historical universe: deletions, tracking rollout, and retained documents condition the result; it is not an estimated market survival rate. Kelly added in September lacks a mature 30-day cohort; its short ages among retired listings do not prove higher liquidity.

Method. Count active priced IDs by merchant_slug; divide the total of the top five by all active priced bag IDs in the snapshot.

Caveat. Share of BirkinBagStock coverage, not sales share or global market share. The same physical bag may appear in multiple listings; stated availability does not guarantee executable inventory.

Method. Join standalone lots and sales results_lots; exclude 22 audited titles unrelated to handbags/fakes/ambiguous items. Closing year; non-null hammer indicates sold. Sell-through = sold / offered; median of 100×hammer/estimate_low among sold lots with positive low estimate, always in the same native currency.

Caveat. Observed sample of multiple Hermès models, not solely Birkin nor all auctions in the market. Auctionet predominates and one Drouot lot is included. Hammer excludes buyer's premium and other costs. Early years do not reach n≥15; 2026 is partial. Ratios do not require FX and do not prove systematically low estimates.

Method. Compare gold/palladium medians by size in standard Togo Birkin; use hardware field or explicit title cues. Exclude unknown or ambiguous hardware.

Caveat. Hardware inferred from titles, without physical verification. Color, condition, age, and resellers vary; does not represent a causal hardware premium.

Method. In standard Togo Birkin by size, neutrals = black/noir, gold, etoupe, and etain; brights = explicit colors recognized by the classifier. Compare medians between recognized groups, without assigning every non-neutral to bright.

Caveat. There are 3,573 unknown colors among active priced IDs. Partial classifier coverage, condition, hardware, year, and reseller mix preclude attributing causality to color; not all shades are comparable.

Method. Divide the current standard Togo median by the benchmark published by the media outlet and subtract one; this is neither a transaction nor an active official quote.

Caveat. Specialized media outlet, not an official Hermès source; benchmark predates the cutoff, before sales tax. Condition, age, and boutique access differ; does not demonstrate realizable profit.