Comment on “Deconstructing Deglobalization: The Future of Trade is in Intermediate Services”

The paper by Baldwin et al. (2024) aims to demonstrate that the future of economic globalization lies in services trade as the expansion of merchandise trade has already begun to run out of steam.

Baldwin et al. begin with a narrative of growth of world merchandise trade relative to world income (gross domestic product [GDP]), both in aggregate and by major trading nations and the three main sectors, manufacturing, mining and fuel, and agriculture. The data suggest that the trade-to-GDP ratio peeked around 2008, even though the peak was not equally present in all major nations and all sectors. However, Baldwin et al. have stopped short of probing the underlying drivers of the observed structural change. If the “hyper globalization” during 1985–2008 had been largely driven by cyclical rather than structural factors, it is hazardous to treat the subsequent slowdown as a structural phenomenon. Moreover, the observed trends in the trade to GDP ratio need to be treated with caution because the well-known structural shift in the composition of GDP from tradable production and toward services. Given this structural change, the measured trade-to-GDP ratio is likely to face downward pressure, thus confirming the fear of deglobalization.

My major concern with the inference of trade slowdown is that it has overlooked possible changes in the price structure of global manufacturing trade associated with production fragmentation (vertical specialization). Production fragmentation essentially means restructuring of the production process of a given product among firms located in different regions and countries. In this process, as Young (1928) postulates, increasing returns take place throughout the industry rather than at the individual firm level, generating cumulative, rather than firm-specific, gains from scale economies. Therefore, goods traded within global manufacturing value chain (GMVCs) could experience slower price increases (and hence faster volume growth) compared to the other traded goods.

There is evidence of significant decline, rooted in developments in semiconductor technology, of prices of information technology (IT) products (computers, and data storage and communicating equipment), automobiles and a wide range of other related GMVC products from about the early 1990s (Jorgenson, 2001; Byrne et al., 2016). The relative stability of the real trade-to-GDP ratio after 2008, compared to the sharp decline in the two nominal ratios in Baldwin et al.'s figure 1, is perhaps indicative of this price effect. However, I suspect that, the price index used to deflate the nominal exports does not fully capture the price lowering effect of global production sharing. An analysis using a new dataset, which is carefully constructed to better capture the price effects of GMVC trade, yields the inference that merchandise trade to GDP ratio has not yet peeked, and has continued to maintain its pre-crisis trend following a sharp dip during 2008–2009 (Athukorala, 2023).

The rest of Baldwin et al. provides a comprehensive analysis of trends and patterns of global services trade based on data pieced together from scattered sources. The key inference that the growth of global service trade has continued to boom is reassuring. I have only two comments, which could help further enrich the analysis.

First, Baldwin et al. have not paid adequate attention to the emerging process of “servicification” of manufacturing: shifting of some manufacturing-related services, which were traditionally treated as part of manufacturing output, to service sector firms as an integral aspect of GMVC operation. Failure to distinguish between these services, which are an integral part of manufacturing, from the intermediate services would artificially increase the role of serves in economic globalization. Also, servicification poses a challenge for trade and industry policy formulation because not only the conventional forms of trade protection but also barriers targeted at services could affect manufacturing (Hoekman & Shepherd, 2017). Therefore, the distinct dichotomy made in Baldwin et al. between services and merchandise trade deserves reconsideration.

Secondly, the data in Baldwin et al. seem to suggest that growth in services trade could further widen the North–South divide in the world economy. Unlike merchandise trade, services trade is heavily concentrated in countries in the North (Baldwin et al.'s figures 7 and 11). Contrary to the remark that export capacity is not a limiting factor for engaging in service trade, the dearth of appropriate skill labor remains a major constraint faced by most developing countries in participating in IT-related modern services. Moreover, modern services are unlikely to match manufacturing in terms of the potential for employment generation, in particular absorbing unskilled labor. For instance, the IT-based services sector in India has a relatively low employment elasticity and employment in that sector requires at least college-level education (Panagariya, 2008, p. 286).