A firm may pay its workers well and give labor a small share of what it makes. There is no contradiction. The wage compares payment with the worker’s time; the share compares the whole payment to labor with the value added by the enterprise. If output per worker rises faster than compensation, the wage can rise while the share falls. The question becomes larger when that enterprise takes business from others. An economy does not count each firm once and average their bargains. It weights them by the activity they command. The rise of the dominant firm can therefore alter the division of income without requiring a common wage cut across the nation. To understand what the superstar has captured, we must look both inside its accounts and at the growing place those accounts occupy in the whole.
Unchanged shares within firms can coexist with a falling share across them. Consider two invented firms, each initially producing one hundred of value added. One pays sixty to labor; the other pays twenty. Together they pay eighty out of two hundred, giving labor forty percent. Now let the second firm expand to three hundred while the first stays at one hundred, and hold each firm’s labor share unchanged. The combined payment becomes one hundred and twenty out of four hundred: thirty percent. Labor receives more money and a smaller fraction. The unweighted average of the two firms’ shares remains forty percent, yet the economy’s share has fallen by ten percentage points. The difference lies entirely in their weights. This example stipulates expansion, without explaining how customers were won, what happened to prices or whether workers gained purchasing power. Its lesson is narrower and decisive: an unchanged division within every firm does not guarantee an unchanged division across them.1
The superstar account makes that movement of weights an economic hypothesis. Autor, Dorn, Katz, Patterson and Van Reenen’s 2020 study found substantial reallocation toward firms with lower labor shares; in manufacturing, it dominated their decomposition of the decline. Their model allows more productive firms to gain business as competition becomes tougher, with a falling aggregate share under particular conditions. This is a route to concentration through success at winning customers. A better distribution system, for example, could spread a fixed administrative staff over more orders, while a rival requires additional clerks as its business expands. Even at the same wage, the expanding enterprise could devote a smaller fraction of its receipts to that work. Establishing such a mechanism requires information about costs and production. Size alone supplies neither a charge of wrongdoing nor an explanation of efficiency.2
A decomposition tells us where a change occurred; it does not, by itself, identify the force that caused it. Autor and his colleagues expressly distinguish their supporting evidence from a definitive causal test. The distinction matters because several histories can leave the same pattern of larger firms and smaller labor shares. A newcomer might displace an incumbent through lower costs; an incumbent might acquire rivals; a network might become more useful as more people join it; a dominant business might obstruct entry. These histories call for different evidence and different judgments. Nor is a national industry necessarily the market in which a particular customer buys or a particular worker seeks employment. A company can spread across many towns while meeting competition in each. The aggregate concentration measure begins the inquiry into power; it cannot settle every local bargain in advance.3
The older temptation is to name everything above marginal cost pure profit. A simple invented account shows why that fails. A firm sells one hundred units at ten each, with constant variable cost of eight per unit and fixed costs of two hundred, including all remaining economic costs for the period. Revenue is one thousand; variable costs are eight hundred; fixed costs consume the remaining two hundred. The price is one and a quarter times marginal cost, but economic profit is zero. The markup describes the next unit, while profit must reckon with the whole enterprise. De Loecker, Eeckhout and Unger therefore examine profitability separately from markups, finding increases in both in their historical data. Their estimates require judgments about production and costs; the accounting distinction survives whichever estimate one accepts. A larger wedge above marginal cost must first pay what marginal cost leaves out.4
Power over the selling price and power over the wage must also be traced separately. A producer may face little competition for its product while competing intensely for workers; another may sell at a competitive price while its employees have few practical alternatives nearby. The first concerns what buyers will pay. The second concerns the terms on which workers can leave, enter or refuse employment. Both can affect labor’s share, but evidence of one does not establish the other. Indeed, the superstar model discussed above assumes competitive factor markets. Its result does not require an employer to suppress wages through monopsony. This is why the distinction between the wage and the share must stay visible to the end: a low share does not establish that the employees receive low wages, and a high wage does not establish that they command a large part of the enterprise’s surplus.5
The answer to the plate’s question is therefore a qualified yes with a specific content. The growing weight of dominant firms can carry income toward enterprises where labor receives less of each unit of value added, and the evidence gives that channel substance. The word capture nevertheless asks for an account of how the change occurred. One must still ask how the firms became dominant, what costs sustain their production, and whether workers’ alternatives have narrowed. Otherwise the account confuses a redistribution of activity with a demonstrated abuse of power, or dismisses a redistribution of income because the winning firms produce well. The two investigations belong together. The superstar changes the share by changing which enterprises make the product; institutions and competition help determine who receives the value those enterprises create.