The phenomenon of offshoring has currently moved to the sidelines of public debate – eclipsed by the financial crisis and deep global recession – but may very well soon return to the policy agenda (Blinder 2009). Signs of recovery are presently much stronger in emerging economies in Asia – most notably in China – than in the US and Europe. In a scenario where growth in Asia takes off while growth in the US and Europe remains sluggish, concerns over offshoring of jobs to low-wage countries are likely to regain momentum on both sides of the Atlantic.
Job characteristics and offshoring
When thinking about how offshoring of jobs affects home employment, it is useful to think of workers in terms of their qualifications and professions. It takes highly trained radiologists to interpret computer-tomography images and X-rays; but some of those skill-intensive tasks are performed offshore today, by US- or EU-trained doctors living in South Asia or Australia through a business practice that has become known as tele-radiology. A janitor’s or doorman’s work, on the other hand, need not require an advanced school degree, but their work can typically not relocate offshore simply because proximity to the maintained facility is indispensable.
In short, a worker’s tasks on the job may crucially determine to what extent offshoring puts employment and earnings at risk. The standard view that globalisation mainly shifts labour demand towards skilled workers and contributes to an increased skill premium in developed countries is not necessarily true when producers offshore intermediate services and bring together workers and electronic equipment at a distance.
Trade in task theory
A recent advance in theory by Grossman and Rossi-Hansberg (2008), who treat offshoring explicitly as trade in tasks, offers a unified way to think about offshoring and its effects on employment and earnings. They show that even if low-skilled jobs are easier to offshore than high-skilled jobs, low-skilled workers may still benefit in absolute as well as relative terms. Offshoring may lower production costs in industries using low-skilled workers intensively. This changes relative profitability and induces a reallocation of workers that involves an increase in the relative wage of low-skilled workers. However, there is still a standard Stolper-Samuelson effect that work in the opposite direction; a fall in the relative price of goods intensive in low-skilled labour mitigates the change in relative profitability and thereby the change in relative wages. In order to understand what would be appropriate policy responses to offshoring, it is thus not sufficient to know whether it mainly leaves high-skilled or low-skilled jobs onshore, but it is also requires knowledge about the direction of general equilibrium effects related to a reallocation of workers.1
A number of task characteristics are potentially relevant for the offshorability of a job – the prevalence of codifiable rather than tacit information to perform the job (Leamer and Storper 2001), the prevalence of routine tasks, especially if they can be summarised in deductive rules (Levy and Murnane 2004), and the job’s requirement of physical contact and geographic proximity (Blinder 2006). In a recent study based on survey answers to questions about job characteristics, Blinder and Krueger (2009) conclude that 25% of all US jobs are potentially offshorable. While this seems like a relatively high number, potentially offshorable is not the same as actually offshored, as rightly pointed out by Alan Blinder in his VoxEU column (Blinder 2009).
While the nature of tasks could perfectly coincide with the skill intensity of a job, so that tasks and skills would amount to the same thing for all empirical purposes, there is no reason for this to be the case. In fact, estimates of the relationship between measures of the offshorability of a job and the skills of the worker holding the job suggest that this relationship is rather weak.2
New evidence from German data
In a study based on information about German multinationals and their German employees, we have analysed the relationship between offshoring taking place within firms and the composition of workers and the tasks they carry out in the German parts of the firms (Becker, Ekholm and Muendler 2009). Multinationals carry out a large part of all trade in intermediate inputs and services and therefore are also major players in offshoring activities.
In our study we use plant data linked to information about employees that allow us to discern tasks, occupations, and workforce skills. We combine detailed occupational information in our MNE sample with survey information on the nature of tasks carried out by workers with different occupations. We focus on two aspects of tasks that may be related to their offshorability:
- Whether they tend to be routine or non-routine.
- Whether they tend to require personal interaction with co-workers and customers or not.
Based on the survey information, we link occupations to the degree to which they involve non-routine and interactive tasks.3 We also use information on educational attainment of workers to create a measure of workforce skills. Moreover, we use the distinction between white-collar and blue-collar workers – a widely used proxy for skills in the literature – to be able to compare our results with previous studies.
We run relatively standard regressions of the wage-bill share of a particular type of worker or task on a measure of offshoring and a set of controls. In order to account for the fact that market-seeking may be a much more important reason for MNEs’ expansion in high-income countries than in low-income countries, we distinguish between offshoring to high-income and low-income countries.
The regressions are run for the following four “advanced” work types:
- non-routine tasks,
- interactive tasks,
- highly educated workers, and
- white-collar workers.
We have a relatively short panel that starts in 1998 and ends in 2001. During this period, overall offshore employment at German multinationals increased by 3.9% in manufacturing and 9.0% in services.
According to our estimates, these changes predict between 2% and 17% of the changes in the wage-bill shares of “advanced” work types at the German plants (Figure 1).
Figure 1. Predicted effects of offshoring at German multinational enterprises
Source: Becker et al. (2009), balanced panel of MNE establishments 1998-2001. Note: Share of observed wage-bill changes predicted by offshore employment at MNEs. Services exclude commerce.
Offshoring thus seems to account for some, but not all, of the changes in the workforce composition at home. We find that offshoring is most strongly associated with a shift towards non-routine tasks in manufacturing (it accounts for 14.1% of the total change) and a shift towards interactive tasks in services (it accounts for 17.1% of the total change).
Offshoring matters for the composition of skills too, but considerably less. It explains about 10% of the increase in the wage-bill share of workers with at least upper-secondary education in manufacturing as well as services and about 11% of the increase in the wage-bill share of white-collar workers in manufacturing. Interestingly, we find no evidence that the relationship between offshoring and the wage-bill share of “advanced” work types differ depending on whether we consider offshoring to low-income or offshoring to high-income countries.
What are the implications of these results for policy? Since our analysis does not capture any general equilibrium effects, we are inclined to say “not many”. However, we do think that there are some important lessons.
- First, at the level of the individual firm, offshoring seems to be associated with a shift towards more educated workers. We also report a positive correlation between the increase in offshoring and the proportion of highly educated workers in the firm. Both these results suggest that offshoring is associated with an increased relative demand for more educated workers, as traditional theories would predict.
- Second, our results support the view that the degree to which jobs involve non-routine and interactive tasks is relevant for their propensity to be offshored.
Still, a large part of the variation over time in the task composition – and in the skill composition as well for that matter – cannot be related to variation in offshoring. There is thus still ample scope for other factors to explain the observed trend towards more educated employees carrying out more advanced tasks in German multinationals.
Note: The views expressed in this column are that of the author alone and do not necessarily reflect opinions shared by Sveriges Riksbank.