The Tide

A factory can cross a border while its owners remain at home. The wages formerly paid in one country are then paid in another, and the profits may return to the same shareholders. This movement gives the question of the share an international dimension. The measure still needs a boundary: the labor share of domestic production compares domestic compensation with domestic value added. Foreign wages enter another country's account, while residents' foreign investment income belongs in the distinction between domestic product and national income. A lost job, a lower price and a larger dividend can arise from the same reorganization and reach different households. Opening to trade can therefore change the size of the product, the location of its production and its division. Each change requires its own explanation before the movement of a national share can be attributed to globalization.¹

Moving a task abroad changes both sides of the domestic share's fraction. In an illustrative business with no intermediate purchases, sales of a hundred finance sixty in domestic compensation and forty in other domestic factor income. Suppose a task that cost twenty in domestic wages is replaced by an imported input costing ten, with sales and every other payment held fixed. Domestic compensation falls to forty. Domestic value added falls to ninety, because the ten paid for the imported input is deducted from sales; the remaining domestic income rises to fifty. Labor's share falls from sixty percent to forty divided by ninety, about forty-four percent. This is a stipulated accounting comparison, with no response by output, prices or displaced workers. It shows a possible route into the statistic. Counting the import payment as domestic value added would describe a different and erroneous fraction.

The productivity gain from obtaining a task more cheaply can also expand demand for the work that remains. Grossman and Rossi-Hansberg make this channel explicit in their 2008 theory of offshoring. Firms combine tasks, some performed domestically and others abroad; making existing offshore tasks cheaper can lower the cost of using the associated domestic factor. Under specified competitive conditions, that gain can raise its wage. Their analysis also identifies relative-price and labor-supply effects, which can offset the productivity effect. Its direction therefore cannot be read from the fact that a task has moved. In the preceding illustration, the ten saved could finance expansion, appear in lower prices or remain in the residual; the fixed-output comparison deliberately leaves those responses out. If compensation and domestic value added instead grew in the same proportion, the domestic labor share would remain constant even as its numerator rose.²

The ability to move production can affect a bargain before production moves. Suppose a firm and its workforce can divide a hundred, the workers can obtain twenty elsewhere, and the firm initially has no alternative return. If they divide the gain above their respective alternatives equally, the workers receive sixty and the firm forty. Now give the firm a credible outside option worth twenty, leaving the workers' alternative and the joint hundred unchanged. Equal division of the remaining gain gives fifty to each. The shift follows from the alternative available on disagreement under the stipulated bargaining rule. A relocation threat that yields nothing after its costs leaves the first calculation intact. Workers may also gain alternatives through trade. The effect depends on whose options improve, by how much, and on the institutions through which those options influence the bargain.³

Adjustment can leave particular workers and places bearing losses that an aggregate gain does not repay to them. Autor, Dorn and Hanson studied changes in American local labor markets over 1990–2007, relating their initial industry specialization to subsequent exposure to Chinese imports. Their 2013 paper used Chinese import growth in other high-income markets to help isolate the supply-driven component, subject to the identification assumptions they state. They reported lower employment and earnings, and greater use of transfer programs, in more exposed local markets. This is evidence about differentiated local outcomes in a particular historical episode. It does not by itself calculate the national labor share or net national welfare. The distinction matters for the person whose job disappears: benefits to consumers or owners elsewhere do not constitute a replacement wage, and the time required to find another livelihood belongs in the account of adjustment.⁴

The evidence directed at the labor share is suggestive about trade while remaining narrower than the whole question. In their 2013 study of the United States, Elsby, Hobijn and Şahin found that industries with greater increases in import exposure tended to have larger declines in payroll shares. They explicitly declined to interpret that cross-industry relationship as causal. They also left the underlying channel open: substitution toward imported tasks and the pressure on industry wage premiums are different explanations. The findings support investigating globalization alongside technology, market structure and institutions; they do not assign every decline in every rich economy to the opening of trade. Nor would a lower share in one country imply an equal increase in another. The composition of production changes in both, and wages abroad can rise while remaining a smaller fraction of the value produced there.⁵

The tide changes the alternatives available to households, firms and states, and its benefits require a distribution as surely as its costs do. Cheaper inputs can enlarge what can be produced; movement can weaken one claimant's position and strengthen another's; displaced work can carry a loss through years of a person's life. A national share gathers part of this history into one fraction. The account of its causes must preserve the people, institutions and geographical boundaries that the fraction compresses. The volume began with the division of a product. It ends with the conditions under which people can claim that product, retain it and take it elsewhere. The next inquiry turns toward the decisions through which they do so: how an available alternative becomes a choice.