Regulating cross-border data flows: Firm-level analysis from Japan

Eiichi Tomiura, Banri Ito, Byeongwoo Kang 12 August 2020

a

A

Cross-border data flows are increasingly critical for modern economies, as Artificial Intelligence (AI), the ‘Internet of Things’, 3D printers, and Big Data gain popularity in many countries. To protect personal data from these new technologies, the EU introduced General Data Protection Regulation (GDPR). On the other hand, emerging countries such as China and Russia impose cyber security regulations restricting cross-border data transmissions for other policy purposes. Goldfarb and Trefler (2019) argue that “data localisation is a privacy policy that could favour domestic firms”.

While it is important to value the protection of privacy, policymakers should also to be wary of the rise of digital protectionism. Based on our unique survey of Japanese firms, we found that only a limited fraction of firms regularly transfer data across national borders or collect data overseas (Tomiura et al. 2019). The mere fact that an overwhelming majority of firms do not transfer data to/from foreign countries, however, does not mean that cross-border data flows are unimportant. Accumulated studies in the field of international economics (especially research on firm heterogeneity in international trade) have established that, while the number of firms that are engaged in global activities (such as exporting or foreign direct investment) is small, they are significantly productive. By linking our survey results with firm-level data derived from official statistics, we characterise firms that regularly transmit data across borders in terms of their productivity. This column reviews our firm-level productivity comparison and discusses its implications for globalization in the digital age.

To collect information on cross-border data flows, in 2019 we distributed a questionnaire to essentially all medium-sized or large firms in manufacturing, wholesale, and information-related service industries in Japan. We obtained responses from 4,227 firms which represents a response rate of 21%. The sample size of this survey is substantially larger than that of a similar firm survey by OECD (2018). Our survey shows that, even if we include firms that collect data without introducing the ‘Internet of Things’, the percentage of firms that regularly collect data overseas is only 11% (Tomiura et al. 2019). While such a finding is informative to some extent, we need to characterise which types of firms are active in cross-border data transmission. For this purpose, we link the survey results with official statistics to collect firm-level data on firm size, productivity, export, foreign direct investment, and other standard corporate variables.1

The comparison of productivity across firms is shown in Figure 1. The surveyed firms are classified into the following three groups: (i) firms that collect data within the home country (Japan) and overseas, (ii) firms that collect data only within the home country, and (iii) firms that do not regularly collect data.2 We measure productivity in terms of labour productivity (per-worker value-added) and ‘Total Factor Productivity’ (defined as a residual from the Cobb-Douglas production function). As demonstrated in the graph, the productivity-premium of firms that are collecting data is clear, irrespective of the productivity measure (especially in the case of firms that collect data overseas). Firms that collect data overseas are, on average, more productive than firms that do not regularly collect data overseas (or in Japan) by 14-18%. We observe the same relationship with firm size in terms of sales and the number of employees. We also confirm the presence of a productivity premium when controlling for industry effects or with quantile regressions.

Figure 1 Productivity comparison of firms grouped by their data collection activities

Note: The productivity of firms not collecting data is normalized to one.

The empirical literature on firm heterogeneity trade models, which began with Bernard and Jensen (1995), indicates that only a limited number of firms are exporters, but that these exporters are substantially more productive than exclusively domestically operating firms.3 In line with this stylised fact, we find that firms that are active in cross-border data transmissions tend not only to be productive, but also globalised. This finding applies not only exporters, but also to foreign direct investment firms, such as multinational enterprises. These findings, when combined, suggest that digital data are intensively transmitted across national borders by a limited number of large-sized, productive, and globalised firms. They are productive enough to cover non-negligible entry costs for cross-border activities, including data transfers across national borders. As the firms that are active in data transfers tend to be large, as well as active in many markets and able to trade with many partners, and possibly also exert wide spillover effects from their superior productivity, we should not underestimate the impact of regulations on cross-border data transfers. From our survey, we also find that the productivity-premium is particularly high among firms reporting that their activities are affected by the regulations on cross-border data transfers. Although our survey does not reveal the value or sensitivity of the data they transmit, it is easy to reason that the impacts of the regulations are likely to be serious, especially if these globalised firms intensively transmit large amount of sensitive data across different jurisdictions.

In the same survey, we also asked firms about their use of 3D printers, which require transmission of design data. Among various technologies recently introduced to manufacturing, 3D printing is among the most important in terms of its potential impact on international trade. As 3D printing essentially makes assembly processes mostly unnecessary, there is set to be a decline in trade in intermediate goods, but an increase in digital data flows.4 However, it is premature to determine the impact of 3D printers on business performance and economic activities alone, as more than 93% of the firms in our sample have not yet introduced 3D printers. We therefore discuss whether the productivity of firms is related to their use of 3D printers. We find that firms that have introduced 3D printers tend to be significantly more productive than those without 3D printers. The productivity-premium of 3D printer users is confirmed even after controlling for industry fixed effects or estimating quantile regression. We also find that, on average, the firms that have introduced 3D printers are large, capital-intensive, R&D-intensive, and are active in exporting and foreign direct investment.

To mitigate the problem of possible reverse causality, we compare productivity levels for both five and ten years ago. The productivity premium of 3D printer users remains confirmed. Figure 2 compares the firms’ productivity and their involvement in global activities in the year 2009, ten years prior to our survey on 3D printers. The figure confirms the gap evident long before the adoption of this new technology. This finding of correlation with pre-dated variables suggests that already-productive and globalised firms tend to introduce 3D printers.

Figure 2 Productivity comparison of firms with and without 3D printers

Note: The productivity of firms without 3D printers is normalized to one. “Exporter” and “MNE” are the share of exporters and that of MNEs among the firms with or without 3D printers.

Cross-border data flows are increasing, and this has led some countries to impose regulations. From our unique survey, we find that firms that regularly collect data overseas tend to be significantly large, productive, and globalised. Our survey also shows that productivity tends to be especially high among firms whose activities are affected by the regulations. These findings indicate that we should seriously evaluate the impacts of regulations on cross-border data flows, even if only a limited number of firms directly transmit data across national borders. Furthermore, the impacts of the regulations are likely to increase in the future, as suggested by the vast, untapped potential of introducing 3D printers.

References

Bernard, A and J Jensen (1995), “Exporters, jobs, and wages in U.S. Manufacturing, 1976–1987”, Bookings Papers on Economic Activity, Microeconomics 26: 67–119.

Goldfarb, A and D Trefler (2019), “Artificial intelligence and international trade”, in Agrawal, A, J Gans and A Goldfarb (eds) The Economics of Artificial Intelligence, Chicago: University of Chicago Press. pp.463-492.

ING (2017), 3D Printing: A Threat to Global Trade.

OECD (2018), “Trade and cross-border data flows”, Working Party of the Trade Committee, TAD/TC/WP(2018)19/FINAL, Paris: OECD.

Tomiura, E (2007), “Foreign outsourcing, exporting, and FDI: A productivity comparison at the firm level”, Journal of International Economics 72: 113–127.

Tomiura, E, B Ito and B Kang (2019), “Effects of regulations on cross-border data flows: Evidence from a survey of Japanese firms”, Discussion Paper No.19-E-088, Research Institute of Economy, Trade, and Industry, Tokyo.

Tomiura, E, B Ito and B Kang (2020), “Characteristics of firms transmitting data across borders: Evidence from Japanese firm-level data”, Discussion Paper No.20-E-048, Research Institute of Economy, Trade, and Industry, Tokyo.

Endnotes

1 The list of firms for our survey basically coincides with that for Basic Survey of Japanese Business Structure and Activity conducted annually by Ministry of Economy, Trade and Industry with legal reporting obligation.

2 Firms that collect data only in foreign countries, despite being exceptions, are counted with firms that collect data both in Japan and overseas.

3 Tomiura (2007) is an early example of firm-level productivity comparison of FDI firms, exporters, foreign outsourcers, and domestic firms in a large sample including small-sized firms.

4 ING (2017) predicts that 3D printing will eliminate almost one quarter of world trade by 2060.

a

A

Topics:  Global economy Productivity and Innovation

Tags:  regulation, technology, data, AI, Japan, trade, productivity, innovation

Professor, Faculty of Economics, Hitotsubashi University; Faculty Fellow, RIETI

Associate Professor of Economics, Aoyama Gakuin University; Fellow, RIETI

Associate Professor, Institute of Innovation Research, Hitotsubashi University

CEPR Policy Research