A hundred leaves the last buyer's hand, and only a portion reaches the person who made the thing. Between them stand the carrier, the warehouse, the wholesaler, the distributor and the shop, each with a claim on the passage. From one end the chain looks like a deduction from the producer's due; from the other, an addition to the customer's bill. Its work is harder to see than its price. Goods must arrive where they are wanted, in the quantities people can buy, at a time they can use them, with someone bearing the interval and answering for a failure. A long chain can perform those tasks well. It can also shelter charges whose explanation is the difficulty of seeing and replacing the hand that takes them. The number of links settles neither case. Each must be compared with what would happen to its work, its cost and its power if the arrangement changed.
The first distinction is between a margin on the books and a profit left after the work is paid for. Take an invented unit purchased from its producer for a hundred, resold by a wholesaler for 120, by a distributor for 150, and by a retailer for 180. The successive gross margins are twenty, thirty and thirty; they add to eighty, exactly the difference between the first price and the last. These are receipts available to meet the intermediaries' expenses as well as reward their owners. If their respective operating costs are twelve, eighteen and twenty, the residuals are eight, twelve and ten. The thirty left across the chain is still before any financing, taxes or other costs excluded from this illustration. Calling the whole eighty profit would erase fifty of work already paid for. Calling the remaining thirty rent would require a further account of capital, risk and competitive alternatives.
The same sequence looks different when expressed in percentages. The wholesaler marks up his purchase by twenty percent, the distributor by twenty-five percent, and the retailer by twenty percent. Applied successively, those factors multiply to 1.8; adding the percentages would give sixty-five percent and miss the actual eighty-percent increase from the producer's price. A margin expressed as a share of sales uses yet another denominator. Nor does multiplying markups explain why they were chosen. It merely describes how the successive prices relate once those prices are given. The mechanism of each firm's decision may depend on its costs, contracts, rivals and customers. A cascade can be recorded by arithmetic without being caused by a universal habit of charging a fixed percentage on whatever the last firm charged.
Even the accounting requires the same unit to be followed through the chain. Suppose a second invented consignment contains a hundred units bought for a hundred in all, but losses leave only eighty units available for sale at two each. Revenue is 160 and the gross difference is sixty, before the costs of handling the consignment. The initial purchase alone now costs 1.25 per unit actually sold. Comparing a one-unit farm price with a two-unit shop price would conceal the lost output and treat the two denominators as though they described identical quantities. Grading, processing and changes in quality create further comparability questions. The practical beginning of an inquiry into a chain is therefore a traceable account of goods, dates, services and losses. The endpoint prices are its invitation, rather than its answer.
The work of a useful link survives its removal from the invoice. A retailer who buys directly from a hundred small producers acquires the need to coordinate deliveries, inspect goods, finance stocks and fill the gaps between shipments. A wholesaler may do those things more cheaply by combining many customers' demands and many producers' supplies. Conversely, a direct arrangement may become preferable when those tasks are easier to perform without a separate firm. The comparison runs through the functions. The recovered Middleman made the central point: grain available in the right place and season is a different service from grain left beside the harvest. Removing a name from the route does not itself transport or preserve anything. A shorter chain earns its economy when the work disappears because it is no longer needed, or is done better by whoever remains.
Coase's account of organization puts the comparison on the same footing on both sides of a firm's boundary. Discovering terms, bargaining, inspecting and settling disputes consume resources. Bringing an activity inside one organization can avoid some of those transactions, while creating its own costs of direction and error. Separate firms can economize through specialization; a larger firm can economize through coordination. Neither ownership form supplies the answer before the work is examined. This matters particularly to a chain, because merger can remove an invoice while leaving a warehouse, a manager and an inventory in place. The visible spread contracts or disappears inside the combined accounts; the underlying resources may be unchanged. To learn whether integration has made the route cheaper, one must follow those resources across the new boundary.
There is a separate loss that can arise even when every physical task is already efficiently performed. If successive firms each have power over price, one firm's markup becomes a cost the next firm takes as given when choosing its own markup. The downstream seller restricts sales in response to an upstream charge that already exceeds the upstream resource cost. In making that decision, it does not receive the upstream profit on an additional unit and therefore does not count that profit among its reasons to sell more. The upstream seller, anticipating the response, makes its own decision. This is the problem of successive markups, commonly called double marginalization. Its premise is pricing power at successive stages under the specified contract, rather than the mere presence of two firms. Competitive links that pass through their costs do not create this mechanism simply by being separate.
An invented two-stage market makes the difference visible. Let the final price be 100 − q when q units are sold, and let production cost twenty per unit. There is one producer and one retailer; the retailer adds no further resource cost, pays a uniform wholesale price w, and chooses the final quantity. At any given w, it maximizes (100 − q − w)q, choosing q = (100 − w)/2. The producer chooses w knowing this response, and maximizes (w − 20)(100 − w)/2. It sets w at sixty. The retailer then sells twenty units at eighty, earning four hundred, while the producer earns eight hundred. Together they earn twelve hundred, but the final price includes the effect of two separate pricing decisions. The retailer treats the forty between production cost and its wholesale bill as a cost of every extra unit it might sell.
A combined owner would sell more without improving a single machine or shipment. With the same demand and the same twenty-unit resource cost, it maximizes (100 − q − 20)q, chooses forty units and charges sixty. Combined profit rises to sixteen hundred, while consumer surplus rises from two hundred to eight hundred. The extra gain comes from coordinating the two pricing decisions; it does not require a technological saving. Even then the combined seller retains monopoly power. A price equal to the twenty-unit marginal cost would permit eighty units, so the integrated forty remains below that competitive benchmark. The example identifies a reason why fewer independent pricing decisions can benefit buyers. It offers no general verdict on a merger that also changes entry, bargaining, rivals' access or operating costs.
A different contract can sometimes coordinate the same decisions while keeping both firms. In the invented market, let the producer charge twenty for each unit and a fixed payment of a thousand for the right to trade. The retailer now chooses forty units because its additional purchase costs the actual twenty; after the fixed payment, it keeps six hundred. The producer keeps the thousand, and buyers face the same sixty-unit price as under integration. Both firms do at least as well as under the original uniform wholesale contract. The fixed payment divides the gain without taxing each extra unit. Its feasibility depends on the simple information, financing and contracting conditions assumed here. With uncertain demand, an unknown retailer or a party unable to finance the fee, the arrangement can become harder to make. The lesson is that the contract between links can matter as much as the number of links.
The distinction between necessary work and restrictive position therefore runs through a chain in more than one way. A distributor can lower delivery costs and still command an excessive return where another distributor is barred from serving the same shops. A retailer can bear genuine stock risk while using control of access to dictate unfavorable terms upstream. A large buyer can also obtain a lower purchase price through better coordination or stronger bargaining, with different consequences for the resources used and the surplus retained by producers. The direction of a price change alone cannot separate these mechanisms. Nor can the invoice tell whether the opportunity to replace a link is real. The test follows the service into an alternative arrangement and asks what prevents that arrangement from being offered on comparable terms.
Opacity helps an unproductive link survive when it prevents the parties from making that comparison. A customer may see only a bundled final bill; a producer may see only the first purchase price; each has too little information to identify where a saving is possible or who could supply it. Better records can reveal duplicated handling, unnecessary delay, a charge without a corresponding task, or a contract that rewards restriction. But complete cost visibility would not by itself create another warehouse, secure entry to a controlled route or give the weaker party a better bargain. Some productive links are obscure because their coordination is difficult to see; some rents are perfectly visible because their owner cannot be bypassed. Transparency makes the chain more intelligible. Contestable alternatives give the information somewhere to act.
The chain earns its length by the work it saves, and its margins must still answer for the trades they prevent.
The answer to the original question is thus found by following both the goods and the decisions. The service behind a margin becomes visible when its work is compared with a feasible replacement, including the reliability and risks that a bare price leaves out. A link can collect beyond that service where its position prevents replacement, and separate useful links can together set terms that restrict the whole chain. Gross margins, net returns, resource savings and gains from better coordination are different measures; the inquiry needs each in its proper place. The next plate turns to the merchant who buys in order to sell, bringing the chain's dispersed functions back into the account of one business. His markup will have to explain the work of carrying a thing to its customer, the risk of holding it until then, and the power to decide what the customer must pay. The argument remains open.