The fluorescence discount is not flat: how 395,554 GIA listings show the price hit running from -23 percent on D to zero on H Medium

Summary

Conventional buying advice treats fluorescence as a single-direction penalty: avoid Medium and Strong fluorescence to dodge a price discount. Across 395,554 active GIA-graded natural rounds in colours D to J and clarities VVS1 to SI1 in the Carat Hunter index, the discount is real but heavily colour-dependent. On colour D, Strong fluorescence costs the seller 16.6% in per-carat price, and Very Strong costs 23.5%. On colour H, Medium fluorescence shows no measurable per-carat discount at all (a 0.2% difference within the noise floor of a 50,000-stone cohort). On colour I, Medium fluorescence comes in 1.5% cheaper per carat than None. The single-number rule of thumb (often quoted as "expect a 10% to 30% discount on fluorescent stones") is wrong by direction at lower colours and overstated at all colours where Strong fluorescence does not actually occur. The cohort confirms what gemologists have argued for two decades: the fluorescence penalty is a function of colour grade, not a fixed industry tax. Data as of 8th March 2026.

What the conventional advice says

Most retailer guides and buyer FAQs handle fluorescence in two paragraphs. The first paragraph describes what fluorescence is (visible blue glow under longwave UV for the most common type, present in roughly 30% of natural diamonds). The second paragraph advises avoiding it. The discount is usually quoted as "5% to 10% for Faint, 10% to 15% for Medium, 15% to 30% for Strong". The advice rarely qualifies by colour grade.

The classic counter-argument has been on the record since the Winter 1997 GIA study on the effect of blue fluorescence on diamond appearance (Moses, Reinitz, Johnson, King, Shigley). The study found that observers could not reliably distinguish fluorescent from non-fluorescent stones in face-up viewing for D to F colours. On lower colours (G to J), Medium and Strong blue fluorescence sometimes improved the perceived face-up colour by counteracting the underlying yellow tint. The trade has known this for 28 years. The retail-side advice has not caught up.

The cohort data provides the missing third paragraph: a per-colour, per-fluorescence-strength price grid showing exactly where the penalty is real, where it is mild, and where it disappears entirely.

The cohort

We pulled every active listing in our index that met our cohort filters. Shape: round only. Lab: GIA only. Origin: natural only. Quality band: colours D through J. Clarity range: VVS1 through SI1. Cut grade: Excellent or Very Good or Ideal. Carat between 0.95 and 2.05ct. Fluorescence labelled None through Very Strong. Final cohort: 395,554 listings.

We computed the median listed price per carat for each colour-fluorescence cell, using listed prices in USD with currency conversion at the daily reference rate. Per-carat price is the cleanest measure here because the fluorescent and non-fluorescent populations within a colour band have slightly different carat distributions. Medium and Strong stones skew marginally larger. Cutters keep more weight on stones that already show fluorescence rather than re-cut to dodge it.

The data: per-carat price by colour and fluorescence

Using "None" as the per-carat baseline within each colour, the discount runs as follows.

Colour D: None $7,036/ct (n=35,102). Faint -3.9% (n=9,396). Medium -10.5% (n=7,488). Strong -16.6% (n=2,930). Very Strong -23.5% (n=642).

Colour E: None $6,376/ct (n=39,387). Faint -0.4% (n=9,701). Medium -5.1% (n=5,644). Strong -15.8% (n=2,184).

Colour F: None $6,575/ct (n=39,170). Faint -10.4% (n=9,854). Medium -11.7% (n=5,335). Strong -23.4% (n=2,148).

Colour G: None $5,374/ct (n=51,389). Faint -5.0% (n=10,823). Medium -10.0% (n=5,156). Strong -15.6% (n=2,193).

Colour H: None $4,497/ct (n=50,179). Faint -1.6% (n=9,383). Medium -0.2% (n=3,963). Strong -8.6% (n=1,686).

Colour I: None $3,976/ct (n=37,202). Faint -9.1% (n=9,402). Medium -1.5% (n=4,184). Strong -18.6% (n=1,941).

Colour J: None $3,474/ct (n=24,772). Faint -8.7% (n=7,732). Medium -3.6% (n=3,697). Strong -10.1% (n=1,819).

The pattern is clear at the colour-band level. On colourless top grades (D-F), the discount is monotonic: each step up in fluorescence intensity costs another 5% to 10%. On near-colourless mid grades (G), the same monotonic pattern holds but at slightly compressed magnitudes. On lower near-colourless grades (H-J), the discount becomes irregular: Faint and Strong cost something but Medium does not, and at H the Medium discount is essentially zero per carat.

Why the discount disappears at H Medium

The mechanism is the one the 1997 GIA study described. On colours G to J, the underlying body colour shows a faint yellow tint visible to a trained observer or to a buyer in the right lighting. Blue fluorescence under daylight (which contains some longwave UV) tints the stone slightly blue, and the additive effect with the underlying yellow shifts the apparent face-up colour back toward neutral. A Medium-fluorescent H reads to most observers as visually equivalent to or slightly whiter than a non-fluorescent H, depending on the strength of the fluorescence and the prevailing light source.

The market prices accordingly. Buyers who would otherwise step up to G or F to escape the H tint can buy H Medium at H Non pricing and get the visual benefit. The supply-demand balance for H Medium therefore mirrors H Non. The discount disappears because the perceived product is not worse.

The same mechanism does not apply at D-F because there is no underlying yellow tint to counteract. Blue fluorescence on a D body simply tints the stone blue, which most buyers (and the trade) regard as cosmetically inferior to a pure-white D Non. The discount on D Strong is therefore real and runs at -16.6% in our cohort.

Three real cert pairs at the boundaries

To anchor the cohort medians with verifiable individual stones.

Colour D, Strong fluorescence: GIA 6362020642 is a 1.00ct natural D VS1 round with Strong fluorescence listed at $6,637. The matched-spec D VS1 round with no fluorescence (GIA 6541715807, 1.05ct) lists at $7,857. The Strong-fluorescent stone lists 15.5% below the non-fluorescent comparator, close to the cohort median discount of 16.6%.

Colour H, Medium fluorescence: GIA 7436781943 is a 1.00ct natural H VS1 round with Medium fluorescence listed at $5,446. The matched-spec H VS1 round with no fluorescence (GIA 5496738291, 1.03ct) lists at $5,080. The Medium-fluorescent stone lists slightly above the non-fluorescent comparator on absolute price, and within rounding on per-carat. This is the cohort-median story for H: Medium fluorescence is a non-event in pricing.

Each cert is verifiable on GIA's report check service. The anchors are illustrative of the cohort centre rather than tail outliers.

How buyers should weight the discount

If your colour target is D-F, the conventional advice is roughly correct. A Faint-fluorescent stone runs 0.4% to 10.4% cheaper per carat depending on the colour band (Faint on E is statistically zero discount, Faint on F costs 10%). Strong runs 16% to 23% cheaper. Buying a Strong-fluorescent D for a saving in the high teens is a defensible trade if you know the visual penalty (slight blue cast in some lighting) and accept it. Most buyers would not trade $1,000 of saving for a visible blue cast, which is why the discount exists.

If your colour target is G-J, the conventional advice is wrong on Medium specifically. A Medium-fluorescent H lists at the same per-carat price as a Non-fluorescent H in the cohort. Buying H Medium gets you the same dollar-per-carat as H Non with a possible (not guaranteed) cosmetic benefit if the fluorescence is blue (rather than yellow or other colours) and your viewing environment includes daylight or other UV-rich light. The 1997 GIA study suggested a face-up improvement is observable in roughly half of cases on G-J Medium; the cohort's pricing parity confirms that the market has internalised this finding.

If your colour target is I-J Strong, the cohort discount is large (-18.6% on I, -10.1% on J). If you accept that the stone may show a slightly blue cast in UV-rich light and a slightly less-yellow cast in normal light, the saving is structural. Buyers shopping at I-J for the budget benefit can extend that benefit further by accepting Strong fluorescence on the same colour grade.

Limitations

The cohort is the universe of currently-listed inventory in our index, not transacted stones. List prices skew slightly above transacted prices, and the negotiation behaviour on fluorescent stones may differ from the non-fluorescent baseline (sellers who already know the stone has a discount may negotiate harder to hold the line). The directional findings (D-F penalty is real, H Medium has no penalty) hold; the precise per-cell percentages may shift by 1 to 3 percentage points in a transactions-only cohort.

Sample sizes vary by cell. The smallest cell reported is colour D Very Strong with 642 stones; the per-carat median there is stable but the 90th-percentile estimate is noisier. Cells with n below 1,000 (mostly Very Strong on E-J) are reported as informational rather than primary findings. The piece's headline claims rest on cells with n above 1,500.

Fluorescence "type" is not in our index at the cohort level. The 1997 GIA study and subsequent gemological work distinguish multiple fluorescence colours. Most fluorescent diamonds glow blue. A minority glow other colours. The cosmetic effect at lower colour grades depends on the fluorescence colour matching the body colour's complementary hue. Our cohort is fluorescence-strength only, not colour. A small fraction of the stones in the Medium and Strong cohorts may show non-blue fluorescence. Those would not produce the cosmetic offset described in the H Medium analysis.

Methodology

Cohort filter spec. Shape = round. Lab = GIA. Origin = natural. is_fancy = false. Colour D through J. Clarity VVS1 through SI1. Carat 0.95 to 2.05ct. Cut in Excellent or Very Good or Ideal. Fluorescence in None through Very Strong. origin_suspect = false. Active listings only. Listed price > $500.

Aggregate cohort: 395,554 listings. Cohort split into 35 cells of 7 colours by 5 fluorescence levels. Per-carat median computed within each cell. Discount expressed as percent change versus the same-colour None cohort.

Cert anchors at four cells: colour D Strong; colour D None; colour H Medium; colour H None. Anchors selected from the cohort centre rather than tail extremes. The cert numbers are verifiable on the issuing lab's public report check.

Cohort medians stamped 8th March 2026. Active listings refresh daily; the per-cell aggregation runs quarterly. For the original gemological context, see the 1997 GIA fluorescence study. For full retailer-inclusion criteria and matching algorithm specification, see the Carat Hunter methodology page.