Date of Award

Summer 6-29-2014

Document Type


Degree Name

Doctor of Philosophy (PhD)



First Advisor

James Markusen

Second Advisor

Keith Maskus

Third Advisor

Wolfgang Keller


Production processes are increasingly fragmented geographically, and the performance of production tasks is spread across countries. As multinational production progresses, an intriguing phenomenon arises which is referred to as countries and firms “moving up the global value chain.” While many people may have an informal understanding of what it means, testable definitions and examinations of the dynamics are lacking.

My dissertation aims to provide a unified framework to address the meaning and mechanism of the dynamics of global production and value chain. Task-based theoretical models are developed to explore and characterize the dynamics, which arise from learning-by-doing. Using firm-level data, empirical support for the important theoretical predictions is found. The model is further extended to incorporate the innovation effect, which explains the rising phenomenon of reshoring.

Following the first and second chapter for introduction and literature review, in Chapter 3, I develop a dynamic task-based model of multinational production. The technology of producing a final good is modeled as a spectrum of tasks ranked by their degree of technological sophistication. The global value chain of an industry is thus described as a sequence of tasks that may be spread across countries, with each task adding value to the final good. “Moving up the global value chain” is then defined as an upgrading in the set of tasks that a country, an industry, or a firm conducts.

The basic model features the critical role of learning-by-doing in the dynamic production process. Initially, developed countries (the North) offshore simple tasks to developing countries (the South). The South may receive tasks beyond its technological capability. By conducting the “beyond” tasks, the South improves its efficiency on relatively sophisticated tasks. This learning-by-doing effect enables more complex tasks to be offshored in the next period. This process continues until the Southern technological capability matches the set of tasks offshored. Both types of countries move up the global value chain during this process -- the South conducts additional and harder tasks, while the North concentrates on fewer but the most highly sophisticated activities. The evolution of multinational production is characterized by the task offshoring threshold moving up to its steady state, with the movement pace slowing down over time -- thus a concave-shaped path. The dynamics of other economic aspects, including wage rates and national welfare, are discussed.

In Chapter 4, I develop the dynamic theory of global production within a monopolistic competition framework. Products are differentiated by variety, with each variety being produced by a multinational firm. Countries and firms move up the global value chain due to firms’ learning-by-doing effects, as the Southern subsidiaries engage in a wider range of and more sophisticated tasks, while the Northern counterparts do fewer but more complex activities. The number of varieties increases and displays a converging pattern of growth during the evolution process, and this expansion at the extensive margin is the main source of welfare gains for both countries.

The situation under autarky and the dynamic gains from offshoring are examined. Under monopolistic competition, the South may experience a welfare loss in the short run upon participating in global production. However, in the long run, the learning-by-doing effect will lead the South to be better off than its autarky situation. Meanwhile, the North enjoys a higher level of welfare at the beginning of joining global production, and the gain continues in the long run. Hence, both types of countries get rewarded from offshoring, though their paths are quite different.

The task-based theory predicts that as multinational production evolves, the Southern country’s share of value added in total value of industrial output increases over time, while the growth rate declines -- thus a concave-shaped path. In Chapter 5, a micro-founded approach is applied to test the dynamics of the value-added ratio (VR) of global production contributed by the South. By using a subsidiary-level dataset on China’s multinational operation spanning 10 years, the evolution pattern of industry-level VR is examined. The results show that convergence evidences are present, and the industrial VR dynamics are mainly driven by changes within multinational subsidiaries.

Chapter 6 extends the model to incorporate innovation in developed home countries. Task allocation depends on countries’ relative efficiency of conducting tasks. When both countries improve domestic technologies simultaneously -- one through learning and the other through innovation, the dynamics of multinational production are determined by the countries' relative speed of technology improvement. Both offshoring expansion and reshoring may occur, where reshoring refers to the phenomenon that previously offshored tasks return to their originating home countries.

Included in

Economics Commons