全球价值链中跨国公司活动测度及其增加值溯源

MNEs’ Activities Measurements and Value-added Tracing in GVCs

跨国公司通过跨境直接投资成立国外分支机构和组建生产网络,是全球价值链的组织者和关键参与者。跨国公司的生产和贸易活动与本土企业具有显著差异,然而现有的全球价值链核算框架却并未对跨国公司与本土企业进行区分。本文在Wang et al.(2017)所提出的全球价值链生产分解框架的基础上,纳入跨国公司异质性,提出了一个能够识别和测度跨国公司活动的全球价值链核算新框架,并识别出传统框架中“被忽视”的FDI相关全球价值链生产活动。基于该框架,本文发现:(1)FDI相关的全球价值链活动约占全球GDP的10%,这意味着传统基于国际贸易测算的全球价值链活动被低估了近乎一半,且这一“低估”问题在高研发强度行业、高收入经济体尤为凸显;(2)与美国相比,中国制造业的贸易相关全球价值链活动占比较高,但FDI相关全球价值链活动则处于相对较低的水平;(3)中国FDI相关的全球价值链活动中,有三分之二以上来自外资企业与本土企业之间的生产关联。此外,本文还考察了跨国公司投资收益的分配情况,在贸易增加值统计方法上实现了从“属地原则”向“属权原则”的转变。在“属权”口径下,中国的贸易顺差和美国的贸易逆差都大幅降低。

Multinational enterprises( MNEs) are central players in the global value chains( GVCs) by establishing foreign branches and forming production networks through cross-border direct investment. While the production and trade activities of MNEs differ significantly from domestic enterprises, the existing GVC accounting framework does not distinguish between MNEs and domestic enterprises.By extending the GVC production decomposition framework proposed by Wang et al.(2017), we take explicit account of the differences between MNEs and local enterprises, and propose a new GVC accounting framework that can identify and measure the production activities of MNEs. In this framework, global production and trade activities are decomposed into five types of value chain activities: domestic value chain activities, those in traditional final goods trade, trade-related GVC activities, FDI-related GVC activities, and GVC activities related to both trade and FDI. The last three are all GVC activities. To reflect the different roles played by MNEs and their branches in the production networks of host countries,FDI-related GVC activities can be further divided into three parts: GVCs among MNEs, GVCs where MNEs are in the upstream link, and GVCs where domestic enterprises are in the upstream link. In addition, we explore the distribution of investment income of MNEs and derive value-added trade statistics based on the ownership-based principle as opposed to the conventional territorial principle.We apply our conceptual framework to the inter-country input-output tables from the OECD Analytical Multinational Enterprises( AMNE) Database that can distinguish between domestic enterprises and MNEs. We produce the share of various types of value chain activities at different aggregation levels( at the global, country, and country-sector levels,respectively), with a particular illustration based on the structural characteristics of FDI-related GVC activities in China.The conclusions of this paper are as follows.(1) The MNE-related GVC activities account for about 10% of global GDP, which roughly doubles the traditional measure of the size of GVC activities.(2) The MNE-related GVC activities are more pronounced in R&D intensive sectors and high-income economies than in medium-low R&D intensive sectors and middle-and low-income economies.(3) Compared with the US, China’s manufacturing industry has a relatively high share of trade-related GVC activities, but a relative low share of FDI-related GVC.(4) More than two-thirds of China’s FDIrelated GVC activities are the results of the production linkages between MNEs and domestic enterprises.In addition, we derive the income distribution matrix of MNEs’ value-added and recalculate the value-added trade of different economies under the ownership-based principle. The results show that MNEs’ cross-border investment income is overwhelmingly captured by high-income economies. In comparison, using the value-added accounting method under the territorial principle to measure income streams attributed to different economies can seriously distort the picture of the GVC income distribution.Using China and the US as an illustration, the US is the main source of cross-border investment, participating in GVCs through outward investment and earning huge investment returns. In comparison, China is the main destination of crossborder investment, participating in GVCs by way of absorbing foreign investment, and remitting a large amount of investment returns to MNEs. Therefore, in terms of value-added trade based on the ownership-based principle, China’s trade surplus and cross-border investment income deficit largely offset each other, and the ownership-based trade imbalance of the US is much lower than official trade statistics. In other words, the large difference in the income balance of crossborder investment between China and the US reflects different roles of the two countries in the GVCs.Both our conceptual framework and the new value-added decomposition data generated that considers the heterogeneity of MNEs can provide new perspectives for future research.