POLICY SPOTLIGHT | How the Policy Change Index Can Help Us Understand China’s Next Moves

When it comes to predicting the behavior of China and other closed regimes, the US intelligence community has consistently failed. Why? Because it has prioritized espionage over other intelligence-gathering methods—despite the counterintelligence capabilities of closed regimes and their ability to apprehend our spies. The Policy Change Index (PCI) can provide a better, “open source” way to understand these regimes. Instead of relying on classified information, foreign informants, and covert operations, we just need to understand how to read between the lines in their newspapers. The PCI does exactly that. By reading the People’s Daily, this machine learning tool can provide a better understanding of the Chinese leadership’s plans than what many China experts can predict.

Watch Their Mouth(piece) 

In its official newspaper, the Chinese Communist Party’s leadership is remarkably transparent about its intentions. But to discover their intentions, you have to know what to look for.

The PCI machine learning algorithm does not tell you what the People’s Daily says. Rather, it tells you what the newspaper says that is markedly different from what it used to say, which serves as a baseline. In other words, the PCI doesn’t measure levels but detects anomalies. Anomalies are what raise flags and demand closer attention from analysts.

The PCI “Beats the Market” in Its China Predictions

During its spring 2019 trade dispute with Beijing, Washington raised tariffs on Chinese imports to pressure the Chinese regime to make structural changes to its economy and move toward a more open system. US policy experts and media sources like the Wall Street Journal said these efforts would succeed. But the PCI was seeing something else.

As the algorithm indicated, Beijing was not preparing the ground for concession by telling its people about the changes the United States wanted to see, such as the protection of intellectual property rights. Instead, it was becoming more hawkish. The PCI suggested we should not expect to see any substantial concessions from China. Two weeks later, the negotiations between Washington and Beijing fell through, and the two sides walked away from each other.

The following year, right after the regime thought it had contained COVID-19 in China, the People’s Daily suddenly began to emphasize the country’s military power. The newspaper had always done so to an extent. But now it was flaunting China’s military capacity more, at levels far above the baseline. The PCI detected this anomaly and flagged it. Was China planning some form of military aggression?

A few months after the PCI identified the new trend, the Chinese began to send fighter jets across the median line of the Taiwan Strait. They were harassing ships in the South China Sea. And they were engaging in confrontations with Indian troops at their shared border. Since then, matters have de-escalated, and for the PCI, things have returned to the baseline, back to “on trend.”

The PCI on China’s Attitude toward the United States

The People’s Daily rarely says anything negative about other countries or their leaders: Chinese propagandists prefer not to mention them. As your mother taught, if you can’t say anything good about a person, don’t say anything at all.

Here, again, we need to pay attention to the trend.

How many times has the People’s Daily used the word “US” over the years? After Beijing joined the World Trade Organization and China was integrating into the global economy, the newspaper increased mentions of the US—for a while. When Xi Jinping became leader, mentions dropped off. Today, the Chinese have little good to say about Uncle Sam. So, mum’s the word. And that’s one trend that continues.

Further Reading 

Zachery Tyson Brown and Carmen A. Medina, “The Declining Market for Secrets: US Spy Agencies Must Adapt to an Open-Source World,” Foreign Affairs, March 9, 2021.

Julian Chan and Weifeng Zhong, “Reading China: Predicting Policy Change with Machine Learning” (AEI Economics Working Paper, No. 2018-11, Washington DC, October 2019).

Julian Chan and Weifeng Zhong, “Will China Fold on Structural Issues? An Algorithm Says Not Any Time Soon,” China Business Review, March 11, 2019.

Julian Chan and Weifeng Zhong, “What Artificial Intelligence Is Telling Us about US-China Relations,” Tribune News Service, August 14, 2020.

Alexander L. George, Propaganda Analysis: A Study of Inferences Made from Nazi Propaganda in World War II (Evanston, Ill: Row, Peterson and Co., 1959).

Guoguang Wu, “Command Communication: The Politics of Editorial Formulation in the People’s Daily,” The China Quarterly 137 (1994): 194-211.