Production is increasingly specialised, with firms concentrating workers on certain tasks that take advantage of outsourced intermediate inputs (Ariu et al. 2019, Timmer et al. 2019). Intermediates are an increasingly large fraction of international trade (Johnson and Noguera 2011, 2017); access to them can increase firm productivity (Amiti and Konings 2007) and it may also lower labour shares in manufacturing (Elsby et al. 2013). Labour share declines have also been attributed to increasing profits due to higher concentration (Barkai 2020) and a shift of production towards larger and less labour-intensive firms (Autor et al. 2020).
In a recent paper (Limão and Xu 2021), we study the relationship between globalisation and specialisation, and its implications for the labour share and income. We define specialisation as the share of intermediates in variable production costs. Figure 1 shows specialisation within manufacturing industries in the US increased about five percentage points between 1987 and 2007, about half of it since China’s WTO entry.
Figure 1 Intermediate cost share and specialisation premium, 1987-2007
Figure 2 shows that the industries with the largest increases in intermediate cost shares over this 20-year period are also the ones with the largest labour share declines, which suggests intermediates are replacing labour (as opposed to capital or energy).
Figure 2 Intermediate and labour cost shares, 1987-2007 change
We develop a framework where the relationships above arise from firms adopting intermediate-intensive technologies as markets expand. A key determinant of adoption is the relative price of labour to intermediates, which captures the trade-off between in-house production and outsourcing intermediates from other firms. In Figure 1 we see that this relative factor price increased at an annual rate of 2.5% and that it is strongly positively correlated with production specialisation. This relative price captures the cost savings from intermediates relative to labour so we call it the specialisation premium and find it is positively correlated with trade openness.
Motivated by these and related findings, we model endogenous firm specialisation. Firms can lower marginal costs by adopting higher intermediate-to-labour intensive technologies by paying a fixed cost. The resulting economy of scale at the firm level captures one aspect of Smith's (1776) specialisation argument: when markets are larger firms have an incentive to specialise their labour into a subset of tasks where they are most productive.
We also incorporate a feedback effect that generates an endogenous input-output multiplier. The increase in productivity from specialisation lowers prices and these products can also be used as intermediates. Thus the initial market expansion lowers the relative price of intermediates, which leads to additional specialisation and demand for intermediates.
Similar to other models with input-output linkages, the input-output multiplier reflects the aggregate intermediate cost share, but a key difference in our framework is that this share is not constant because of endogenous specialisation. We show endogenous specialisation has interesting and important implications. Increases in market size, for example due to lower trade costs, generate:
- larger real income gains relative to exogenous specialisation;
- an increase in the aggregate variable cost share for intermediates and a decrease for labour;
- an increase in concentration of profits and sales, as the most intermediate-intensive firms are the ones that benefit the most from the reduction in the relative price of intermediates.
A calibration of the model to US manufacturing in 1987–2007 illustrates the importance of trade and technology shocks in explaining several facts, including the relative increase in intermediates to labour share. We carry out two policy experiments based on our calibration. First, an industrial policy (e.g. a tax/subsidy) that induces all firms to specialise would have increased real income, so the equilibrium is inefficient (firms don’t internalise the externality of their adoption decision on others). That income increase is significant if the policy was implemented in 1987 but negligible in 2007 since, by the latter period, trade and technical change had induced sufficient specialisation. Second, we compute the impact of an increase in trade costs of 16 log points, similar to the recent trade war, and show it increases the labour share but reduces market size and real income substantially – almost half way to the predicted effect of the US shutting all trade.
Amiti, M and J Konings (2007), “Trade Liberalization, Intermediate Inputs, and Productivity: Evidence from Indonesia”, American Economic Review 97(5):1611–1638.
Ariu, A, K N Hakkala, J B Jensen and S Tamminen (2019), “Service Imports, Workforce Composition, and Firm Performance”, VoxEU.org, 22 November.
Autor, D, D Dorn, L F Katz, C Patterson and J Van Reenen (2020), “The Fall of the Labor Share and the Rise of Superstar Firms”, Quarterly Journal of Economics 135(2):645–709.
Barkai, S (2020), “Declining Labor and Capital Shares”, Journal of Finance 75(5):2421–2463.
Elsby, M W, B Hobijn and A Sahin (2013), “The Decline of the US Labor Share”, Brookings Papers on Economic Activity 2013(2):1–63.
Johnson, R C and G Noguera (2011), “The value-added content of trade”, VoxEU.org, 07 June.
Johnson, R C and G Noguera (2017), “A Portrait of Trade in Value-added over Four Decades”, Review of Economics and Statistics 99(5): 896-911.
Limão, N and Y Xu (2021), “Size, Trade, Technology and the Division of Labor”, NBER Working Paper 28969.
Smith, A (1776), An Inquiry into the Nature and Causes of the Wealth of Nations, Volume 1, Strahan and Cadell, London, UK.
Timmer, M, S Miroudot and G De Vries (2019), “What Countries Do When Exporting: Measuring Functional Specialization in Trade”, VoxEU.org, 01 February.