The cut, seen from both sides
A reader meets this chapter’s question from whichever side of the counter they happen to stand on. The seller’s version: whoever has sold an app, a jar of honey, an hour of freelance work, or a season’s harvest has watched a platform, a shop, or a buyer take a slice that dwarfs what the making earned. The buyer’s version: the eggs cost one price at the farm gate and three times that price six miles away, and nothing about the eggs changed on the drive. Either way the suspicion lands in the same place. Somewhere between the person who makes and the person who pays sits a person who neither makes nor pays, and that person seems to keep the largest share. The suspicion is old enough to have medieval law named after it: English market codes treated buying up goods on their way to market as offenses, forestalling and regrating, on exactly the theory that the man in the middle earns by standing there.
The previous chapter answered the other half of the question people usually ask in one breath, why there is sometimes only one seller to buy from, and it ended with a promise: that the middleman’s cut has less to do with villainy than with who bears risk, who owns the customer, and who solves the matching problem. This chapter pays that promise. Its claim, stated up front: margin accrues to functions, and the functions that command margin are risk-bearing, matching, and the customer relationship. Production commands little, however honest the sweat, because production is usually the step with the most willing replacements. That logic does not care what is being made, and the chapter’s last section follows it to the place most readers feel it personally, the market where what is being sold is your labor.
The ladder from the farm gate
The claim needs a real chain, walked end to end, and coffee has one of the few honestly documented ones. In 2002, researchers from Oxfam traced a kilogram of Ugandan robusta from a farm in Kituntu district to a jar of instant coffee on a UK supermarket shelf, interviewing every link on the way and recording the price at each hand-off. The farmer selling unprocessed beans got 14 US cents for the kilogram, in green-bean terms. The local trader who trucked it took 5 cents; milling and its haulage took another 5; bagging and the ride to Kampala, 2. The exporter’s processing, taxes, and margin added 9 cents, the journey to an Indian Ocean port 10, ocean freight and insurance 7, and the importer’s landing and delivery charges 11, bringing the beans to the roaster’s door at 63 cents a kilogram. Instant coffee concentrates weight, about 2.6 kilograms of beans per kilogram of soluble powder, so the bean cost inside a retail kilogram was about $1.64. That kilogram sold in the UK for an average of $26.40.
Two things about the ladder deserve more attention than the bottom rung usually gets. First, the margins in the middle are thin. The trader, the miller, the exporter, and the shipper together account for 49 cents a kilogram, and most of that is cost rather than profit; the managing director of one Kampala export house told the researchers that exporters were content to clear a single US cent per kilogram, and that some grades were not worth the diesel to move. Competition does to middlemen in open stretches of a chain exactly what the previous chapter says it does wherever entry is cheap: it shaves the margin toward the cost of doing the work. Second, the money is not in the middle. It is at the far end, in the $24.76 that appears after the beans reach the roaster, the block that pays for roasting, freeze-drying, packaging, advertising, shelf space, and the retailer’s and roaster’s margins. A local coffee buyer for one of the trading houses put the resulting bafflement to the researchers plainly: “Are the roasters cheating? Are they making superhuman profits?” The chapter’s work is to replace that question with a mechanism.
What the middle is paid for
Start with the function that is easiest to see once drawn and almost invisible otherwise. A market with no intermediary is not a market where buyers and sellers meet freely; it is a market where every buyer must separately find, vet, and haggle with every seller. The arithmetic of that is brutal. Connect 4 growers to 4 shops directly and someone must maintain 16 relationships; put one trader in the middle and 8 suffice. Scale the numbers up to a real market and the middleman stops looking like a toll booth and starts looking like a piece of infrastructure: one counterparty who knows the growers, grades the crop, holds the stock, and can be found in the same place every day. Economists who model this, Ariel Rubinstein and Asher Wolinsky among the first, treat the middleman as a seller of search itself: buyers and sellers pay the spread because finding each other unassisted costs more time than the spread costs money. Trust rides on the same rail. A dealer who trades every day has a reputation to protect and the expertise to grade quality, which is why used-car lots, art dealers, and coffee graders exist in every economy that has used cars, art, or coffee; the one-off seller has neither (Daniel Spulber’s survey of intermediation, and Gary Biglaiser’s model of middlemen as experts, carry the theory).
The second function is risk. Between the farm gate and the shelf, somebody owns the coffee at every moment, and whoever owns it eats whatever happens to it: the warehouse fire, the wet season, the forty-percent slump in world prices between one harvest and the next. The maker’s exposure ends at the hand-off; the intermediary’s begins there, multiplied by inventory. A wholesaler holding three months of stock in a falling market loses money on goods it has already paid for, which is a species of loss the maker never sees, and part of every spread is the insurance premium the chain charges itself for carrying that exposure. The third function is the one that explains the fat end of the coffee ladder: ownership of the customer. The roaster and the retailer hold the two assets scarcest in the whole chain, a brand the shopper reaches for without thinking, and the shelf the reaching happens at. Growing coffee is something hundreds of thousands of farms on four continents can do, and the price of what many can supply gets driven toward cost. Holding the loyalty of a nation of breakfast drinkers is something a handful of roasting houses and retail chains in each market have achieved, and what few can supply collects what the many cannot.
The replaceable step
That asymmetry, many who can make, few who hold the gate, is the chapter’s actual answer, and it is worth stating as a rule: the share of the final price a link in the chain captures tracks how hard that link is to replace, and almost nothing else. Not effort. Not hours. Not virtue, and not even how essential the step is, since every step is essential; remove any one and there is no cup. The farmer works hardest and captures least, because the farmer competes with every other farmer on earth, while the gatekeepers compete with the two or three others who share the gate. Where a middle link sits in an open, contestable stretch of the chain, the Kampala exporters clearing a cent a kilogram, its cut stays honest for the reasons the previous chapter gave; where a link has built or inherited a moat, scale, a network, switching costs, or a friendly rulebook, the cut grows to whatever the position will bear. The middleman problem and the monopoly problem are the same problem wearing different clothes.
Nothing tests the rule like the newest middlemen, the ones made of software. The platforms were sold, in their springtime, as the end of the middleman: the store with no shopkeeper, the taxi rank with no dispatcher, everyone connected to everyone. What they became is the strongest intermediaries in economic history, because they industrialized all three functions at once: matching by algorithm, trust by ratings ledger, and ownership of the customer so complete that the customer carries the store in a pocket. The toll reflects it. Apple’s App Store takes a 30% commission on digital sales under its standard terms (15% for developers under a million dollars a year), rates that regulators and lawsuits have already carved down in some jurisdictions, which is to say that the most valuable shelf ever built prices like the scarce asset it is. The developer supplies the app; so do two million others. How platforms that charge the user nothing settled on attention as the thing they retail is Volume II: History’s chapter on how attention became a product; the moats that keep their position from eroding are Chapter 5’s and are not restated here.
When the thing you make is your labor
The rule was stated with goods, and it does not care about the distinction. Most readers make exactly one product, their working hours, and sell it through chains with middles of their own. Some of the intermediaries are visible and priced: the staffing agency that bills the client one rate and pays the temp another, keeping the spread for matching and for carrying the payroll risk; the gig platform that owns the rider’s customers the way the roaster owns the breakfast drinker; the recruiter paid a percentage of the first year’s salary for solving a search problem. Others are so structural they stop looking like middlemen at all. An employer is, among other things, an intermediary between a worker and the customers the worker will never meet, bundling labor with capital and a brand and bearing the risk that the product does not sell; part of why wages are steadier than revenues is that the firm keeps a spread in good years to eat the bad ones. The uncomfortable corollary of this chapter’s rule follows the worker too: pay tracks replaceability before it tracks contribution. Why that is, and what credentials, licenses, and the option to walk away do to it, is the opening argument of Volume IV: Work, in the chapter on what makes a job good or bad and why pay tracks leverage rather than honor, which builds on the reading given here.
The same lens prices a fact this book returns to in Volume V: being poor means buying more of life through intermediaries, and through the expensive kind. The check-casher, the payday lender, the rent-to-own shop, and the money-transfer counter are all middlemen selling matching, advance, and trust to customers the cheaper gates exclude, and their fees compound into a premium on poverty itself. That ledger is itemized in Volume V: Pathologies, in the chapter on why it is expensive to be poor; it belongs in this chapter only as evidence that the middleman’s spread is not an exotic business phenomenon but a line item running through ordinary lives, thickest where the customer has the least leverage.
Cutting out the middleman
The oldest promise in commerce is the one in the shop window: buy direct, cut out the middleman, keep the difference. Sometimes it delivers; the farmer selling at a Saturday market keeps the retail margin she used to surrender. The instructive part is what she gets with it: the stall fee, the van, the four unpaid hours of standing, the unsold crates driven home, the customers found one at a time. The margin did not disappear, and was not stolen back; it was reunited with the work it was always paying for. She has not escaped retailing, she has become the retailer, and whether that trade pays depends on whether she does retail work more cheaply than the shop did. The same accounting follows every direct-to-consumer brand that proudly skips the store and then pays an advertising platform to be findable: the shelf was a middleman, the search box is a middleman, and the acquisition cost moved from one line of the budget to another. Functions can be moved, merged, automated, and re-priced. What they cannot be is skipped, and whoever performs them where they are scarcest will keep the biggest cut, whatever the era calls that person.
So the answer the reader can carry out of this chapter is compact. The middleman captures more than the maker wherever the middle functions, matching, risk, and the customer, are scarcer than the making, which in a connected world of many able makers is most of the time; the cut turns predatory only where the middle has moats, and then the right diagnosis is the previous chapter’s, not a morality tale about middlemen. What this chapter priced across a value chain, the next two price across time. The next chapter asks why whole categories of the reader’s budget, the hospital, the classroom, the daycare, grow relentlessly dearer while televisions collapse in price; the answer starts with what this chapter noticed about labor, and it is the volume’s second great case of a price pattern that looks like villainy and runs on arithmetic.