IMAGE: GUY BERRESFORD
THE RELATIVE price of materials in the United States rose sharply in the 2000s. Over the same period the share of national income going to workers fell. A new working paper by Juanma Castro-Vincenzi and Benny Kleinman argues that this is no coincidence. Their finding, rooted in a careful study of U.S. manufacturing industries and local labour markets, has unsettling implications for trade-dependent economies like Nepal, where leaping import bills and a depreciating currency have pushed up the cost of intermediate goods. If that relationship holds, higher materials prices may be silently eating into workers’ take-home share of the economic pie.
The logic is deceptively simple. When firms buy materials, they combine them with labour and capital to produce output. The labour share is conventionally measured out of value added—gross output minus materials expenditure. A rise in materials prices, the authors show, raises the share of production costs absorbed by materials. If firms have positive profits and materials are complements to primary inputs, this cost reallocation leaves a smaller slice of value added for labour and capital, while the profit share of value added rises. No increase in markups or changes in returns to scale are required. The effect is mechanical but potent.
The authors test the idea in several ways and find the same result each time. Across 361 U.S. manufacturing industries between 1991 and 2016, a 1% bump in materials prices slashed labour’s share of value added by about 0.23%. Similar results appear across U.S. regions, when using different sources of price shocks, and in 23 other countries. The effect was strongest in the United States and Japan, where materials prices rose the most, and weaker in Europe, where exchange rates cushioned the increase. Their estimates suggest that without the rise in materials prices, the decline in labour’s share during the 2000s would have been about one-quarter to one-third smaller. Higher materials costs also reduced investment, reinforcing the effect.
For Nepal, the implications are worth considering. The country runs a persistent trade deficit and imports a wide range of intermediate goods, from petroleum products to machinery and chemicals. The rupee has depreciated against the dollar in recent years, yanking up the local-currency cost of those imports. If the paper’s mechanism applies, the effect on workers’ share of income may be non-trivial. The authors note that the transmission of higher materials prices to the labour share is stronger under higher ex-ante profit shares and higher materials intensity. Manufacturing, where profits are often concentrated and materials costs loom large, would be the most exposed sector. As import bills soar, the profit share of value added may increase, even as the absolute level of profits could decline.
The paper does not claim that materials prices are the only driver of labour share movements. Changes in markups, factor-biased technical change and shifts in the capital-labour mix also matter.
But the authors estimate that materials-price fluctuations account for a big share of the medium-term variance in the labour share, and for around a third of the decline in the 2000s. Other explanations, such as rising concentration or automation, are not mutually exclusive. They are, in the authors’ framing, complementary forces that operate through distinct channels.
The policy implication for a country like Nepal is that rising import costs are not merely a drag on the trade balance. They may also be shifting the distribution of income away from labour and towards profits, without any change in market power or business conduct.
And that the mechanism is structural rather than cyclical. It operates as long as materials and primary inputs remain complements and firms earn positive profits—conditions that are unlikely to change in the near term. For a government concerned about worker welfare, the paper offers a sobering reminder: the price of imported inputs matters not just for inflation but for who gets paid, too. ■
