Offshoring components and their effect on employment: firms deciding about how and where

Abstract

Firms must take two fundamental decisions: how and where to produce. Traditional measures of offshoring include information on both decisions but cannot distinguish between them. In this article, we attempt to distinguish the evolution of the requirement of inputs per unit of output (how to produce) from the delocalization of production to others countries (where to produce). We call global technical change to the first element and net offshoring to the second. We further decompose net offshoring into net inter-industry substitution and intra-industrial offshoring (replacement of domestic inputs for imported ones from the same sector). This last measure quantifies better the concept of delocalization of production to other countries looking for lower costs, the original idea behind offshoring. This decomposition allows us to further investigate on whether technical change or net offshoring is the main factor in recent Spanish industrial employment changes.

Notes

¹ We are only aware of one study by Görg and Hanley (2005) that studies the effect of offshoring on employment but at firm-level and using survey rather than input–output data.

² The original term employed by Feenstra and Hanson was outsourcing; however, recent literature about the topic has moved to the analysis of offshoring, a concept which comprises a firm imported purchases, independently of whether the provider is financially linked or not of the purchaser, while outsourcing refers to firm purchases to other independent firms, which could be domestic or foreign (for further analysis, see Díaz-Mora, 2008).

³ See Feenstra and Hanson (1996, p. 244).

⁴ This technical change assumes that each imported coefficient increases at the same rate than the average of total intermediate inputs.

⁵ From the assumption of firms maximizing profits in a perfect competition environment, it is possible to obtain the demand function for the labour factor from the first-order condition, which states that each factor's marginal product has to equal its real price.

⁶ To calculate different offshoring measures, we employ the use matrices of the Spanish input–output tables, instead of the inter-industry symmetrical (commodity by commodity) matrices. Our decision is justified by data availability for the period 1993 to 2002, as we have at our disposal six usetables (1995–2000) for one symmetrical table, which allow us to take into account changes in the table coefficients. Also, we measure offshoring directly from the use matrices. It is possible to observe a very important change in the coefficients from 1995 to 2000. This is why, to fill the gaps we estimate the data for 1993 and 1994 by extrapolating the growth rates of 1995 to 1998, and for 2001 and 2002 we apply the growth rates of 1998 to 2000.