The Changing Nature Of The Trade Landscape Calls For Open-Source Intelligence

The Russian invasion of Ukraine has made Chinese military action against Taiwan seem less abstract and heightened interest in the potential economic fallout of a war in the South China Sea. Traditional methods in the economist’s toolkit — computational general equilibrium (CGE) models and econometric analysis — are the gold standard for analyzing trade deals and even sanctions. But they are often inadequate to anticipate the enormity of an unusual events or major conflict.

Helpfully, necessity is the mother of invention, and the growing availability of big data, including open-source intelligence, offers new facets of study.

The current economic analytical toolkit largely revolves around CGE models and econometric analysis. These tools presume we have many precedents and plenty of data representing them and are relatively precise with small deviations from the status quo. But what happens when we face international economic and trade policy questions that we haven’t fully faced before?

For the past few decades, arguably since the end of the Cold War, we have been living in a one-model world. The nature of international trade and economic questions have largely revolved around the themes of liberalization and deregulation. When it comes to those “standard” issues, CGE models have been especially good for contemplating hypotheticals and econometrics for understanding the past.

CGE modeling, a tool of choice for U.S. trade policy analysis, is usually used for ex ante (“before the event” in Latin) questions; that is, the potential effects of a proposed policy. The U.S. International Trade Commission, which is a go-to for independent trade and economic analysis by the House Ways and Means Committee, the Senate Finance Committee, and the U.S. Trade Representative, has used CGE since the early 1990s. They have tackled such questions as “What are the the potential economic effects of a US-UK free trade agreement?” and “What is the likely impact of the US-Mexico-Canada Agreement?” As computational power has increased over the years, these models have gotten more detailed, and can drill down to activity across hundreds of sectors and countries, and even at the subnational (e.g., state) level.

For ex post (“after the event”) questions, econometrics is the most popular approach. The method is used to look back and examine a whole host of events and policy changes, like the long-run labor market effects of the Canada-U.S. Free Trade Agreement, the productivity effects of increased foreign direct investment in Mexico, the effects of a natural disaster on global value chains, and trade and inequality.

Each tool has its limitations, even in the one-model world. There are entire chapters in trade agreements such as digital trade, electronic commerce, state-owned enterprises, and competition policy which are hard to capture in CGE models. Even among the most sophisticated econometric analysis, it can sometimes be hard to disentangle correlation from causation.

But the one-model world seems to have passed. New geopolitical conflicts, a pandemic, the Ukraine war, Brexit, rising populism and unilateral trade actions, and now China’s increasingly aggressive posture in the Indo-Pacific all have implications for international economic policy. It is hard to characterize any of these disruptions, or potential disruptions, as small deviations from the status quo. We are dealing with unusual events more often these days, and the current toolkit is lacking.

The changing nature of the policy landscape makes this a good time to add something new to the toolbox. That’s where open-source intelligence and big data come in. (Big data include data that are conventional and unconventional, such as text, satellite images, videos, multimedia files, audios, etc.) I think economists interested in the empirical international trade analysis of large disruptions should take note.

Consider a potential Chinese invasion of Taiwan. Where does one even begin to assess its potential economic impact? Much of that depends on what the invasion looks like and which parts of the global economy are vulnerable to a kinetic conflict in the Taiwan Strait and nearby waters.

In a recent policy brief, my colleague Weifeng Zhong and I attempt to address some of that using an unusual open-source data set: a collection of points of interest in Taiwan with detailed coordinates, curated by a malicious Chinese entity. The data suggest that the kind of military planning China may have for Taiwan potentially includes transportation facilities like seaports and information and communications technology facilities like submarine cable landing stations, where subsea cables, the backbone of the world wide web, come to shore.

We argue that a Chinese invasion of Taiwan would not only severely disrupt container shipments in the Taiwan Strait and nearby waters, but also could also knock the island off-grid in the digital economy and break critical links in global value chains, putting high-tech sectors like semiconductor manufacturing in jeopardy. If addressing the scenario with standard CGE modeling, one might see a more formalized look at tariff equivalents or negative productivity shocks — but the real-world version would likely be so upending that even the effects of the most punishing tariffs or productivity strikes would be no match for those of invasion.

The further we wade into uncertain times, the more frequently economists will be called upon to provide information and analysis around rare events. Sometimes, the questions are less about how big the economic effects will be, and more about what the nature of the shock will be. This is where new methods like open-source intelligence and big data are most needed.