Sectoral productivity, intermediate inputs, and structural transformation

This thesis is a collection of three essays. All share the similar motivation of highlighting the importance of the intersection between sectoral productivity growth and intermediate input composition.

The first paper, entitled ”Recent Trends in Labour Productivity and TFP Growth in the U.S.”, provides a basis to demonstrate the contributions of accounting for intermediate inputs to explain the slowdown in sectoral labour productivity growth and TFP in the U.S. since the 1980s. First, I perform a growth accounting exercise in which I decompose labour productivity growth into the contributions from capital, labour, intermediate inputs and TFP, finding the latter to be the main driver of the slowdown. Of novelty, I use input-output tables to disaggregate the contribution of services inputs towards these productivity terms. Based on this, I explore the efficiency gains and losses in labour productivity and TFP from the reallocation of inputs. On the sectoral labour productivities, I calculate the contributions from the respective subsectors, finding small efficiency losses from the movement of workers towards less productive industries. I implement a similar approach to examine the changes in sectoral TFP growth, deducing that the efficiency variations from the reallocation of inputs can account for most of the sectoral slowdown, particularly from changes in the composition of services inputs. More specifically, both goods and services sectors have experienced large but differential contributions from the professional business services subsector since the Great Recession, as the calculations show that there were TFP efficiency losses for the goods sector, and efficiency gains for the services sector. These results show that the sectoral productivity growth measures can be influenced significantly by the intermediate input composition in each sector.

In the second chapter, titled ”Intermediate Input Elasticities and Biased Technical Change”, I explore differences in sectoral production functions by estimating the sectoral elasticity of substitution across intermediate inputs, and their associated input-specific technical rates. Using U.S. industry-level data during the period 1987-2019, I estimate relative input demand functions using a supply-shock instrumental variable to control for endogeneity. I find that services industries combine material and services inputs as substitutes, while goods industries combine them as complements. In addition, I estimate the direction of technical change to be services input-augmenting across most industries. Furthermore, I estimate sectoral elasticities and technical rates between intermediates and value added, finding that both goods and services industries combine them as substitutes. These results point towards the existence of technical bias towards using materials more efficiently in the goods sector, while technology is biased towards value added in the services industries. Alternative approaches that estimate the elasticity between intermediates and labour yield results supporting the substitutability estimates with the intermediate input bundle, however indicating instead labour-biased technical change in the goods sector, and services input-bias in the services sector.

Lastly, the third essay with the title ”Intermediate Input Technologies and Structural Change” studies the drivers of input expenditure shares across sectors and production factors in the U.S. during the postwar period (1950-2019). Based on the estimation results from the previous chapter, I present a model where there is input-specific technologies and sectoral differences in the elasticity of substitution. Drawing on key equations of the production side, I infer these technologies such that they can reproduce the data on relative prices, input shares and relative sectoral output. I assess the magnitude of neutral, sector- and input-specific technical change and study their consequences for the patterns of intermediate input expenditure by using a number of counterfactuals. My findings indicate a major role for the resulting intermediate services input-biased technical change in explaining the current trend of services input-deepening across sectors, because of the combination of different elasticities between intermediate inputs and services-augmenting technical change. In addition, these reallocations occur while also showing that neutral technology growth in the goods sector is the main driver of the process of structural transformation towards services.