Causal Inference in International Political Economy: Hurdles and Advancements

Causal Inference in International Political Economy: Hurdles and Advancements

Causal inference is gaining significant momentum in social science fields, including economics, political science, and sociology. Causal thinking is revolutionary on multiple dimensions: scholars have not only become more attentive to understanding “X causes Y, and why?”, but have applied innovative methodological tools to advance research of important social phenomena, including natural, survey, lab, field, and elite experiments, among others.

In particular, with the groundbreaking publication of Designing Social Inquiry: Scientific Inference in Qualitative Research in 1994, political scientists have endeavored to apply quantitative and qualitative methods to explain “the logic of inference” (King, Keohane, and Verba, 1994). However, experiments are less frequently utilized in top journals of International Political Economy (IPE), a subfield of Political Science, compared to other fields, such as American Politics, Comparative Politics, or even Economics. IPE scholars are interested in important global economy questions, such as the flows of money and investment across state borders, domestic origins of international trade preference, multinational firms’ political and economic activities, effectiveness and problems of international standards and organizations, and so on. Why is causal inference used less frequently in IPE? And how should IPE studies move forward?

An Enduring Methodological Battle: Statistics vs. Experiments

One of the conundrums lies in the methodological battle between correlation (which statistical tools are good at) and causation (which experiments are designed for). IPE scholars used to rely heavily on cross-national data and/or qualitative case studies to generate theories, but these approaches are not immune to problems. Early researchers were alerted to a fatal yet deceptive mistake: correlation does not imply causation. 

That being said, it is often tempting to argue that social problems or policy outcomes are the direct results of certain social factors or policy interventions, although they sometimes just happen to be “correlated” with the outcome. This issue is then in fundamental contradiction with the spirit of causal inference, in which “X” is thereason, or one of the reasons, for the emergence or development of “Y.”

But why aren’t experiments a pervasive methodological tool in IPE? The problem of randomization is the key. Unlike other fields, the common units of analysis in International Relations (IR) are international and regional organizations, states, governmental agencies, interest groups, private firms, and non-profit organizations. It is a natural headache for researchers to randomize a fairly large group, such as a state or a multinational firm. 

What’s more, scholars need credible counterfactuals to evaluate the impacts of a policy instrument, but seeking a proper counterfactual for a historical event is empirically challenging. For instance, the CHIPS and Science Act was introduced by President Biden in 2022 to protect American economic resilience and national security from external adversary forces; yet, we do not know what would have happened if such an Act had not been introduced. Scholars also viewed the experiments on state elites as “too risky, too costly, or too difficult to implement” (Dietrich, Hardt, and Swedlund, 2021), given that it is often hard and costly to reach out to political elites and invite them to join an experiment.

So far, readers may wonder if the main problem is the “international” component of IPE - it is difficult to randomize large units and locate credible counterfactuals, and more importantly, generating a grand theory about the international system seems way too distant from micro-level experimental tests on individuals and small groups. However, in recent years IPE scholars have dedicated efforts to explore new ways of applying experiments in IPE studies. Let me briefly guide you through some outstanding examples below.

Advancements in IPE Research: Survey, Field, and Lab Experiments

First, survey experiments are so far the most commonly applied experimental approach in IPE research. Due to the advantage (and feasibility) of manipulating individuals and small social groups with a specific survey instrument, survey experiments infer how public opinion works in real-world settings (Gaines, Kuklinski, and Quirk, 2007). By randomly assigning targeted people to receive different information prompts, survey experiments aid researchers in mimicking what happens in our societies. For instance, Steinberg and Tan (2022) investigate how protectionism adopted by a foreign government affects the home country's public attitudes toward free trade. They fielded three original survey experiments in China and primed people in the treatment group about U.S. protectionist measures, whereas leaving the control group without any information clues about them. They further replicated the results in Argentina.

Moreover, IPE scholars are very innovative in designing and implementing field experiments. To understand the effectiveness of international transparency law and standards in anti-money laundering, Findley, Nielson, and Sharman (2013) confidentially incorporate 1,264 corporate service providers from 182 countries. In particular, they utilized email aliases and pretended as “international consultants” to reach out to companies, after getting approval from the Institutional Review Board. Later on, Daniel L. Nielson collaborated with many colleagues and filed several field experiments to test important and interesting topics, such as electoral cycles and investment incentives (Jensen, Findley, and Nielson, 2020), and foreign aid preferences from the mass public (Milner, Nielson, and Findley, 2016). 

What’s more, frontier IR studies have also increasingly focused on political elites as the targets for lab experiments. As Hyde (2015) notes in a review article, inviting elites to participate in lab experiments is very relevant to IR theories, ranging from foreign policy decision-making processes (e.g., Moxnes and Van der Heijden (2003)) to gender differences in formulating a policy stance (e.g., McDermott and Cowden (2001)). However, most of these articles are related to foreign policy and international security, rather than international political economy. Further, as Hafner-Burton et al. (2013) argue, a realistic problem is that “experienced elites are difficult to obtain…and (are) skittish about revealing information about their decision-making processes and particular choices.” Hence, figuring out how to best utilize the resources of political elites and the relevance of using them in generating new IPE theories (or testing existing ones) become critical questions for future scholars. 

As we move forward, it will become more important to think about different research methodologies as a whole, whether they are qualitative methods, statistical methods, experiments, or computational social sciences tools (see here for a past D-Lab Blog Post on computational social sciences). These approaches should be deemed complementary to each other, rather than replaceable. Ultimately, future scholars need to be considerate about how causal inference can aid in the advancement of IPE theories, as well as the strengths, weaknesses, relevance, and applicability of each research methodology in each project.

References

1. Dietrich, Simone, Heidi Hardt, and Haley J. Swedlund. 2021. “How to make elite experiments work in International Relations.” European Journal of International Relations 27 (2): 596-621.

2. Findley, Michael G., Daniel L. Nielson, and J.C. Sharman. 2013. “Using Field Experiments in International Relations: A Randomized Study of Anonymous Incorporation.” International Organization 67 (4): 657-693.

3. Gaines, Brian J., James H. Kuklinski, and Paul J. Quirk. 2007. “The Logic of the Survey Experiment Reexamined.” Political Analysis 15 (1): 1–20.

4. Hafner-Burton, Emilie M., D. Alex Hughes, and David G. Victor. 2013. “The Cognitive Revolution and the Political Psychology of Elite Decision Making.” Perspectives on Politics 11 (2): 368-386.

5. Hyde, Susan. 2015. “Experiments in International Relations: Lab, Survey, and Field.” Annual Review of Political Science 18: 403-424.

6. Jensen, Nathan M., Michael G. Findley, and Daniel L. Nielson. 2020. “Electoral Institutions and Electoral Cycles in Investment Incentives: A Field Experiment on Over 3,000 U.S. Municipalities.” American Journal of Political Science 64 (4): 807-822.

7. King, Gary, Robert Keohane, Sidney Verba. 1994. Designing Social Inquiry. Scientific Inference in Qualitative Research. Princeton, NJ: Princeton University Press.

8. McDermott, Rose, and Jonathan A. Cowden. 2001. “The Effects of Uncertainty and Sex in a Crisis Simulation Game.” International Interactions 27 (4): 353-380.

9. Milner, Nielson, and Findley. 2016. “Citizen preferences and public goods: comparing preferences for foreign aid and government programs in Uganda.” The Review of International Organizations 11: 219-245.

10. Moxnes, Erling, and Eline van der Heijden. 2003. “The Effect of Leadership in a Public Bad Experiment.” The Journal of Conflict Resolution 47 (6): 773-795.