The Asynchronicity Of Retail and Wholesale Prices

Price negotiations for pre-owned watches and jewelry tend to be far more inefficient and messy than those for diamonds, where the Rap Sheet sets a starting base price, and the gist of negotiations is generally about future price expectations. The bid-ask messiness of finished luxury goods such as watches/jewelry vs luxury commodities such as diamonds is usually attributed to finished goods’ more complex, heterogenous nature (ie differences in condition/materials/brand/more components/etc).

But that is not the whole story.

In this post I argue that in addition to heterogeneity of watches/jewelry, the lack of a centralized wholesale trading platform is a significant cause of market inefficiency, as it enables and often exacerbates the asynchronicity of wholesale and retail prices.

Firstly, note that explicit, long-term supply contracts do not exist in our industry because we are a secondary market of heterogenous product. Although there are long standing relationships between wholesalers and their customers (retailers, other wholesale dealers, etc.), each “deal” is negotiated separately, resembling retail. Hence there are no guardrails to wholesale price fluctuation. There are also no common industry metrics for valuation such as ‘price per unit’ or ‘yield %’ to benchmark.

On the contrary, primarily due to the difference in pace of wholesale and retail buying, wholesale prices are more flexible and forward looking than retail prices, which overall tend to be stickier. Just to be clear, retail pricing is generally cost-plus; I am not saying that wholesale price levels are higher than retail price levels (although surprisingly, they sometimes are).

Here is a real, recent example. A novice buyer approaches my booth at the IWJG show and is interested in rarer watch A, which I offer at $4500. He then whips out Chrono24 and finds a listing for a different specimen of watch A listed for $5200. Upon this discovery, he says, “Well if $5200 is retail on Chrono24, I should be buying it for $3500! Chrono24 takes fees! Your price is practically retail!”

In this common example, the novice buyer assumes that the sellers of comparable listings should have similar sourcing costs, and/or the buyer should be offered a perceived wholesale price to compensate for liquidity risk (ie the 33% discount). And if not, the seller (me) should lower profitability or take a loss. The basis for this assumption is retail listing data from Chrono24.

Sourcing is generally a deal-by-deal, forward looking economic activity, and sensitive to market demand and supply; Because dealers usually only source watches that are commonly understood to increase in value, the sourcing cost (wholesale price) will tend to increase at a faster pace than retail over time, resulting in retail price lag or stickiness. Of course, there are many cases where retail price leads wholesale price (usually rarer collectors’, vintage or trendsetters’ watches), but generally, wholesale prices increase at a faster rate than retail. The reverse is also true; wholesale prices tend to decrease faster than retail - but in practice, professional buyers rarely source watches on a downward trend.

Retail listing inertia¹ (ie the $5200 listing the novice buyer found on Chrono24) gives a false impression of price stability, while upward price pressure builds up beneath, unbeknownst to a buyer watching retail price movements. The buyer erroneously assumes that supply is available at a perceived discount to that retail price, when in reality the wholesale price level has moved up. This asynchronicity is most probably due to the pace of buying - dealers are forced to compete with each other to buy inventory regularly to continue their business, pushing prices up faster; consumers/collectors buy at a more leisurely pace, driven by choice, not necessity.

In the example above, the novice buyer visited my booth twice during the show, bewildered at my reluctance to come down on price for watch A. 5 days later, I sold watch A at the same $4500 price to a different buyer who in turn sold it within 30 days for $6300 retail, a well-deserved 40% margin and a 21% increase to the $5200 Chrono24 listing that the novice buyer showed me. And that Chrono24 listing? Gone. Today I would only be able to source watch A at a hair below the price I sold it for. Retail lag or stickiness is a major source of confusion for all participants in this market. This is also usually why data analytics based on retail (and auction) prices tend to over- and under-shoot and are not very reliable forward price indicators.

Here is another, opposite example. Some participants in our market will say that for example ‘Brand Z watches are trending because the auction hammer prices for Brand Z models increased by X% compared to the last auction season.’ Some then go on to say, ‘Buy Brand Z now or you’ll never be able to afford one!’ while others will say ‘Brand Z is so overhyped! Sell before they crash!’ Still others, ‘Brand Z watches are timeless!”

These claims are similar to what

Scott Sumner

calls reasoning from a price change - it does not explain why the price changed, whether the increase is due to higher demand, supply constraints, temporary hype or a combination of all? We also do not know by whom the Brand Z watch was bid up, the extent of due diligence (for example physical check or online photos only) by the bidders, the cross-price elasticity between models of the brand, and most importantly, the depth and composition of the pool of buyers. It is a single data point in time, untethered to real-time wholesale markets, making it an unreliable forward (or even present value) indicator. Even condition, which many professionals swear by, does not always have a linear relationship with auction hammer price. Hence auction watchers and buyers often find themselves unable to reconcile the gaps between the all-in hammer price, the retail Chrono24 price, and achievable resale price when it comes time to sell. Granted, consumer/collector preferences are nebulous and the industry fragmentation makes it difficult to gauge supply (I tried to explain Cartier’s rise here, and here ).

Now let’s go back to my previous example. What is the relationship between retail and wholesale price and how should it be calculated? The correct way to answer this question is to develop a pricing model based on past transaction data. Unfortunately, extreme heterogeneity, market fragmentation and limited public transaction data currently make these models error-prone and sometimes even misleading when pricing individual watches.

Instead, here is a very simple formula that models the relationship between wholesale and retail prices based on 4 factors:

\(\small \begin{aligned} P_r = C_w \times \Bigg(1 &+ k_1 \left(\frac{1}{S_m} - 1\right) \\ &+ k_2 \times G \\ &+ k_3 \left(\frac{1}{T_m} - 1\right) \\ &+ k_4 \times P_s \Bigg) \end{aligned}\)

Where:

\({\small \begin{align*} P_r &= \text{Retail Price} \\ C_w &= \text{Wholesale Cost} \\ S_m &= \text{Supply Multiple} \\ G &= \text{Expected Growth Rate in Price} \\ T_m &= \text{Turnover Multiple} \\ P_s &= \text{Prestige of the Seller (scale from 0 to 1)} \\ k_1, k_2, k_3, k_4 &= \text{Weighting Constants for Each Factor} \end{align*} }\)

Sm and Tm are multiples, so 1 = baseline or unchanged, >1 means increased supply or faster turnover, and <1 means reduced supply or slower turnover. (Prestige of Seller and Turnover would overlap if solving for a specific client, so should be adjusted accordingly.)

The formula assumes that the relationship between the wholesale and retail price is a function of:

1. supply

2. turnover speed

3. expected price growth

4. prestige (brand) of the seller

I did not include watch condition, overhaul/repair cost, brand, etc. because this would already be accounted for in the wholesale price (or retail if backing into the wholesale price). Going back to Watch A which I sold, I assume the following:

• Pr = Retail Price

• Cw = $4500

• Sm = 0.8

• G = 0.015

• Tm = 1

• Ps = 0.6

• k1,k2,k3,k4 = 0.5

For simplicity, I assigned equal weights (0.5) to all factors, reflecting an assumption that each contributes equally to price movements in the absence of a regression model to determine data-driven weights. I used 6-month CPI for the growth rate, a slightly above average 0.6 seller prestige, and kept turnover speed as unchanged. I assumed supply decreased slightly by 0.2x, which is what prompted me to source it in the first place. Plugging those assumptions into the formula, I get a Retail Price of $6446 for Watch A, close to what my retailer customer sold it for.

Although by no means perfect, this is probably a better way to think about pricing rather than arbitrarily trying to back into a wholesale price by discounting from sticky retail prices (like the novice buyer above). While ‘Prestige of seller’ is quite subjective, some watch price data providers such as Subdial compile and publish (retail) Turnover and Supply multiples. Growth rate is also subjective, but I would just use past 6-month CPI. Many people would deride this formula for not including physical characteristics such as condition, but as aforementioned I assumed it would be priced in the wholesale price and did not want to double-count it as a factor. The purpose of the formula is to better understand the relationship between wholesale and retail, not forecast or compute price.

Experienced wholesalers have a better pulse on actual supply and turnover and hence are quicker to react. You will find them touring booths at tradeshows counting supply (like counting cards in blackjack) while recalling their retail customers’ turnover. It’s almost as if we are subconsciously running the above formula in our heads. However, because the market is globally fragmented (the IWJG tradeshow is not the only venue), and human data carrying capacity is limited, inefficiencies remain.

A centralized, global online wholesale trading platform would help mitigate these inefficiencies by providing real-time wholesale pricing, turnover and supply data, improving price discovery, and enabling wholesalers and retailers to react more synchronously.

Unfortunately, data is lacking to backup my claims, and due to the fragmented nature of our industry, my own data may not be representative of other wholesalers. So I am relegated to give you theories backed up only by anecdata. I would be grateful if you could prove my theory wrong with data, as it is a source of much frustration for me during negotiations. These days when someone quotes Chrono24 or auction prices I start zoning out. Writing these essays is not my main job, and I don’t have the time to compile data on the asynchronicity of wholesale and retail prices - if anybody is interested in taking this on (doubtful), I’m ready to support you financially.

1. Retailers tend not to adjust pricing unless its sales/discount season. And some retail listings remain unchanged due to genuine stagnation in demand, rather than outdated pricing. But empirically, buyers have a pretty good idea of what is/will be in demand and select accordingly.