Fine Wine Index: How the Main Market Indices Are Built
Updated
There is no such thing as the fine wine index. There are five in common use, they are built from three different prices, and they disagree with each other in the same month.
A fine wine index is a basket of wines tracked from a fixed starting value, usually 100, so what you read is movement rather than price. What separates one index from the next is the price fed into the basket. Liv-ex calculates its Fine Wine 50, 100 and 1000 from the Mid Price, which its index documentation describes as "the mid point between the current highest bid price and lowest offer price on the Liv-ex trading platform". Cult Wines builds its Global Index from Wine-Searcher data, calling it "a dynamic set of wines that represents the fine wine investment market based on strict liquidity criteria". Wine Market Journal builds its indices from auction results. A wholesale bid, a merchant asking price and a hammer price are three different numbers for the same bottle, so three indices covering the same wines move at different speeds and by different amounts.
What is a fine wine index, and what does the number mean?
It is a level, not a price. Liv-ex says the Fine Wine 1000 "is calculated monthly and was rebased at 100 in December 2003", so its reading of 350.7 on the Liv-ex indices page means the basket is worth roughly three and a half times its December 2003 value. It does not mean a case costs £350.
That makes index levels useless to compare across families unless you check the base first. Two indices can both read 300 and describe different histories if one starts in 2003 and the other, like the Cult Wines Global Index with its "Inception date as of 1 January 2014", starts eleven years later, according to CultX.
The spread within a single publisher is the more useful lesson. Liv-ex's index page in early August 2026 showed the Burgundy 150 at 611.6 and the Champagne 50 at 500.7, against the Rhone 100 at 174.6 and the Port 50 at 151.4. One publisher, one exchange, four answers to "how has fine wine done". If your cellar is Rhône and Port, the headline benchmark has been describing somebody else's wine.
Which fine wine indices actually matter?
Five families cover almost everything you will be shown in a pitch deck.
| Index family | What is in the basket | Price it uses | Update |
|---|---|---|---|
| Liv-ex Fine Wine 50 | the last ten 'physical' vintages of the five Bordeaux First Growths, each represented by "one 12x75cl case" | Liv-ex Mid Price | Daily |
| Liv-ex Fine Wine 100 | "100 of the most sought-after fine wines for which there is a strong secondary market" | Liv-ex Mid Price | Monthly |
| Liv-ex Fine Wine 1000 | "1,000 wines from across the world" in seven regional sub-indices, including the Bordeaux 500, Burgundy 150, Champagne 50, Rhone 100 and Italy 100 | Liv-ex Mid Price | Monthly |
| Cult Wines Global Index | Nearly 7,000 wines across seven regional indices | Wine-Searcher data | Weighted monthly, "rebalanced annually based on strict liquidity criteria" |
| Wine Market Journal (TOP500, HIVAL, WMJ150, MATURE) | Auction lots, from a database the publisher says covers virtually all trades of rare wine at live and internet auctions in the world | Realised auction prices | Not stated |
Knight Frank does not build its own. Its 2026 Luxury Investment Index reports the wine line using the Liv-ex Fine Wine 100, which it says "posted a decline of 2.5% in 2025", with "total losses at almost 25% since the 2022 peak". So the luxury-asset league table you see quoted in the press is the Liv-ex basket wearing a different hat.
One footnote worth carrying into any comparison. Liv-ex's index page lists a Bordeaux Legends 40 and a Rest of the World 60, while its Fine Wine 1000 description lists a Bordeaux Legends 50 and a Rest of the World 50. Constituent counts move. Check which version of a basket a chart is drawn from before you set two charts side by side.
Which price goes into the index, and why does it matter?
Because the three candidate prices are not the same number, and the gaps between them are where your money lives.
Liv-ex publishes two. The Mid Price sits between live bids and live offers on the exchange and, in Liv-ex's own words, is "the most robust measure for pricing wines available in the market". Liv-ex defines the Market Price differently: "the best listed selling price for a standard case in the secondary market", sourced from wine businesses on and off the exchange and "widely used for valuation statements as a guide to replacement cost". Liv-ex uses the Mid Price for its indices and the Market Price for its Power 100 brand ranking. One publisher, two price series, two purposes.
Then there is the venue. Wine Market Journal quantified this in November 2025: across its TOP500 and HIVAL indices, wines sold roughly 0.4% higher online than at live auction, while its WMJ150 blue-chip index showed a 4% premium at live sales and its MATURE index of 1980s and 1990s vintages ran 12% higher live. Twelve percent, on the same wines, decided by which room the lot sat in.
An index that samples one venue, or one wholesale exchange, is not wrong. It is narrow. Liv-ex describes itself as "the global marketplace for the wine trade", and its Fine Wine 100 constituents must be "physically available in the UK market". That is a trade-to-trade, UK-physical price. It is not what a private buyer in Singapore pays after premium, duty and shipping.
How do wines get into the basket, and how do they leave?
By committee, on published rules, and the leaving is the part nobody models.
Liv-ex sets three tests for the Fine Wine 100. A wine needs "critical acclaim from a leading critic (a 95-point score or above)", it needs a regular market on Liv-ex, and it must be "physically available in the UK market", which excludes wines still only offered en primeur. Constituents are then "reviewed on a quarterly basis by committee", according to Liv-ex.
The Fine Wine 50 instead runs on a mechanical rule, according to Liv-ex. Components are added in July as vintages become physical and removed after ten years, so the basket is always the last ten physical vintages of Château Lafite Rothschild and its four peers. Most sub-indices of the Fine Wine 1000 use the same rolling window, from "the ten most recently physical vintages for 50 top Bordeaux chateaux" down to "the ten most recently physical vintages for five Southern and five Northern Rhone wines". The exception is the Bordeaux Legends, "a selection of 50 Bordeaux wines from exceptional older vintages (from 1982)".
Read those rules back and you see what an index cannot contain. A wine that stops trading regularly leaves. A vintage that ages out of the ten-year window leaves. A wine that never earned 95 points never arrives. The basket is a permanent selection of the liquid and the acclaimed, refreshed four times a year, which is a fair description of the market's best-behaved corner and a poor description of a real cellar.
How is a fine wine index weighted?
By scarcity, not by equal shares. Liv-ex multiplies each wine's Mid Price by "the wine's average production level - such as 20,000 nine litre cases - with this figure gradually reduced as the wine ages", on the logic that drinking shrinks the float.
The Fine Wine 1000 then stacks its regions. At launch the weights were Bordeaux 500 at 46%, Bordeaux Legends at 22%, Burgundy 150 at 14%, Italy 100 at 7%, Rhone 100 and Rest of the World at 4% each, and Champagne 50 at 3%, according to Liv-ex. Liv-ex's 1,000-wine index was, on the day it started, more than two-thirds Bordeaux.
That matters for anyone whose holdings sit in the Burgundy producer atlas rather than on the Left Bank. A 3% Champagne sleeve cannot move a 1,000-wine index no matter what Champagne does. Cult Wines takes the other route and weights its regional indices "dynamically weighted each month to accurately reflect the amount of activity on Wine-Searcher", which tracks attention rather than production. Neither is neutral. Weighting is an opinion about what the market is, expressed as arithmetic.
Why does a fine wine index understate the risk?
Because a wine that did not trade is carried at its last price, and a flat line looks like calm.
Fogarty and Sadler set this out in the Journal of Wine Economics in 2014, testing six ways of measuring wine returns on 14,102 auction sales of investment-grade Australian fine wine. They describe the commercial index approach, in which "wines are held at their last sale value until they sell again", and they are blunt about the consequence: the assumption "means that the approach is likely to understate the risk associated with wine investment". The measurement is stark. In Fogarty and Sadler's figures, on quarterly returns, the commercial index approach produced a standard deviation of 2.99%, while the hybrid model they treat as the reference produced 6.04%. The commercial index halved the apparent volatility of the same wines over the same period.
Masset and Weisskopf reached the same conclusion from the other direction in Economic Modelling in 2018, in a paper on wine indices they subtitled "Nicely labeled but slightly corked". Their finding: "both the volatility and the beta of fine wine is understated when estimated with existing wine index data. The true volatility and beta of the First Growths from Bordeaux appear to be close to 20%, respectively 0.45-0.60, suggesting that the diversification potential of fine wine is more limited than commonly believed."
That is the sentence to carry into any conversation about wine as portfolio ballast. The low correlation and the smooth line are partly a property of the measuring instrument. Our guide to wine investment risks takes the same point through to what it does to a real holding.
Do different methods give different answers on the same wines?
Materially, yes. Fogarty and Sadler found that the choice of estimation method has a significant impact on the estimated return distribution, with mean quarterly returns ranging from 1.64% under the hedonic model to 2.23% under the average adjacent period method on identical data.
The direction of each bias is predictable once you see the mechanism. Repeat-sales indices track only wines that sold twice, and since "higher quality vintages trade more frequently" there is a potential sample selection problem with this approach. Fogarty and Sadler concluded that, as in the art market, "repeat sale approaches overstate returns in the wine market".
Two more properties nobody advertises. Regression-built indices are revised, according to Fogarty and Sadler: "after each auction the entire index needs to be re-estimated, and, hence, after each auction there are revisions to historical estimates of the return to wine", so last year's chart is not the chart you saw last year. And the start date does real work. Dropping the first five quarters of their sample raised the mean quarterly return by around half a percentage point across all six methods. When a marketing chart begins at a flattering date, that is not a coincidence, it is the largest free variable the author had.
Is the Liv-ex Power 100 a fine wine index?
No. It is a brand ranking, and it uses a different price series from the indices.
Liv-ex publishes the method in full. In the 2021 edition, according to Liv-ex, brands had to have "traded at least three wines or vintages, and had a total trade value of at least £10,000" over a twelve-month window, and were then "ranked using four criteria: year-on-year price performance (based on the Market Price for a case of wine on October 1st 2020 with its market price on September 30th 2021); trading performance on Liv-ex (by value and volume); number of wines and vintages traded; and average price of the wines in a brand". Each criterion carried a weighting of 1, except trading performance at 1.5, "because it combined two criteria".
So a brand can climb the Power 100 on trading activity while its prices go nowhere, and Dom Pérignon placing highly tells you the trade moved a lot of it, not what a case returned. Useful for reading liquidity. Not a price series, and not a substitute for one.
What an index cannot tell you about your own bottles
It cannot tell you what you own is doing. An index is a basket assembled on somebody else's rules: liquid wines, 95 points and above, UK physical, ten vintages deep, rebalanced quarterly. Your cellar is a specific list of wines in specific vintages at specific condition, and the gap between the two is the number that decides whether you sell.
See the benchmark and your own wines on one chart
You get the comparison the index families do not offer: our market index next to the individual wines you hold, built from realised auction observations, with the number of lots behind every point shown so you can see where the sample thins out.
That is the difference between knowing fine wine fell last year and knowing whether yours did. Set up a comparison, watch the wines you actually own against the basket, and decide on the gap rather than on the headline.
Start with the full method behind our scoring and price data if you want to check our arithmetic before you trust it, the investment guide if the question is whether to buy at all, and drink-now if the honest answer for a given case is that it was never an investment.