University of Michigan - Ann Arbor

Establishing causality in economics is a significant challenge for two main reasons. First, unlike in many natural sciences, it is often impossible to conduct controlled experiments to isolate a single variable. The economy is a dynamic system where numerous factors change simultaneously, creating confounding variables that can lead to misleading correlations. Second, the economy's fundamental complexity, arising from the interactions of millions of people, makes it impossible to measure and understand every variable. This inherent complexity means economists must be skeptical of simple claims that one complex phenomenon, like capitalism, is the direct cause of another, such as a rise in living standards. To address these challenges, economists use methods like natural experiments and advanced statistical techniques to infer causal relationships from observational data.

The Challenge of Establishing Causality in Economics

Which of the following best describes a challenge in establishing causality in economics?

What is a key reason why controlled experiments are rarely feasible in economics?

Which method can economists use to explore causal relationships when controlled experiments are not feasible?

Why is it difficult to establish causality in economics through experiments?

In a video discussion, economists Thomas Piketty and James Heckman explain how collecting data is fundamental to their research on economic inequality and the development of policies to reduce it. Heckman argues that data is essential for testing and differentiating between competing economic theories, a capability that earlier economists like Adam Smith lacked despite their profound ideas. Piketty reinforces this by cautioning against developing overly sophisticated theories that are not grounded in empirical facts, advocating for a more systematic collection of data, especially historical data, to enable proper comparisons and inform public debate. Both economists agree that this data-driven approach is critical for understanding complex issues like economic inequality and for designing effective policies. Piketty also suggests that while data is not a cure-all for social conflict, it is a prerequisite for a more rational and informed public discourse.

Piketty and Heckman on Why Data is Fundamental for Inequality Research and Policy

The emergence of capitalist institutions, which can foster prosperity through specialization and technological innovation, coincided with the start of the continuous technological revolution and the 'hockey stick' growth in living standards. However, this temporal correlation does not definitively prove that capitalism caused this economic takeoff. A major confounding factor is the Enlightenment, a significant cultural shift towards free-thinking that also emerged in the 17th century, around the same time. Consequently, economists and historians debate whether the primary driver of this sustained growth was the new economic institutions, the cultural shift, or a combination of these and other factors.

Debating the Causes of Hockey Stick Growth: Institutions vs. Culture

A typical example of a causal statement in economics is the claim that a reduction in the central bank's interest rate will lead to an increase in consumer purchases of homes and cars. This statement posits a direct cause-and-effect relationship, where the policy action of lowering the interest rate is the cause, and the resulting change in consumer behavior is the effect.

Example of a Causal Statement in Economics

A natural experiment is an empirical study where a naturally occurring external event, such as a change in institutions or a natural disaster, is used to investigate causal relationships. Researchers compare the outcomes for the economic actors who were affected by this event (the treatment group) against those who were not (the control group). This scenario mimics a controlled experiment because the external event creates distinct comparison groups without the researcher's intervention, allowing for causal inferences about the event's impact.

Natural Experiment

To illustrate a common problem in establishing causality, consider a study on the effect of merit-based scholarships on future income. A simple comparison between the earnings of scholarship recipients and non-recipients can be misleading. This is because students who earn the scholarship likely possess higher inherent ability, a factor that itself contributes to higher future wages. This 'higher ability' is a confounding variable, making it difficult to isolate the true impact of the scholarship from the pre-existing advantages of the students who receive it.

Confounding Variables in Scholarship Studies

Based on the principles of establishing economic causality, critique the following statement and explain why the conclusion might be flawed.

Evaluating an Economic Causal Claim

A government implements a new job training program for unemployed individuals. One year later, a report finds that 60% of the program's participants have found full-time employment. Explain why this statistic alone is insufficient to prove that the training program *caused* this employment outcome.

Interpreting Policy Outcomes

An economist observes that cities with a higher number of libraries per capita also tend to have higher average incomes. Explain why this observation alone is insufficient to conclude that building more libraries will cause a city's average income to rise. In your answer, describe a potential alternative explanation for this relationship.

Analyzing an Observed Economic Relationship

A city government introduces a new subsidy for citizens who use public transportation. A year later, official data shows a 5% reduction in air pollution in the city center. The mayor's office issues a press release stating, 'Our public transport subsidy has directly caused a significant improvement in our city's air quality.' Critically evaluate this claim. In your response, identify at least two other possible factors that could have contributed to the improved air quality and explain why their existence makes it difficult to definitively prove the subsidy was the sole cause.

Critiquing an Economic Policy Claim

A government official announces: 'One year ago, we implemented a new job training program. Today, the national unemployment rate has fallen by 1%. Therefore, our program was a success and caused this decrease in unemployment.' Based on the challenges of establishing cause-and-effect relationships in economics, provide two distinct reasons why this conclusion might be flawed.

Evaluating a Causal Economic Statement

Economists often cannot use traditional controlled experiments to establish causality. Match each proposed economic study with the primary reason why such an experiment would be unfeasible or inappropriate.

A correlation is a statistical assessment that two variables have moved together, whereas causation implies that a change in one variable directly produces a change in the other. Establishing a causal relationship is more restrictive because it requires identifying the underlying mechanism that connects the variables. Therefore, simply observing that two things have moved together is not sufficient proof of a cause-and-effect link.

Distinguishing Correlation from Causation

Conducting controlled experiments in economics faces significant hurdles. It is often ethically problematic and practically unfeasible to manipulate major economic variables for large groups of people. For instance, a government cannot randomly assign half its population to a new tax system while keeping the other half as a control group. Similarly, experimenting with interest rates or unemployment levels on a national scale is not possible. These constraints force economists to develop alternative methods, such as natural experiments and statistical techniques, to infer causality from observational data.

Challenges of Controlled Experiments in Economics

Economists aim to establish causal relationships to achieve two primary goals: to understand the fundamental reasons behind economic events and to develop effective policies for improving the economy's performance.

The Purpose of Causal Statements in Economics

An economy consists of the countless interactions of millions of individuals, making it an extraordinarily complex system. Because it is impossible to measure and comprehend all of these interactions, economists must be cautious about making simple causal claims. Asserting that one complex system, such as capitalism, directly causes a specific outcome, like increased living standards or environmental problems, oversimplifies the intricate reality. This inherent complexity necessitates skepticism towards definitive cause-and-effect statements about large-scale economic phenomena.

Complexity of Economic Systems and Causal Claims

Controlled experiments, which are the scientific standard for establishing causality, are generally not feasible for answering the most significant economic questions. This practical impossibility requires economists to find alternative ways to study cause and effect, such as by analyzing real-world observations through methods like natural experiments.

Limitations of Conventional Experiments in Economics

The scientific standard for establishing that one variable (X) causes another (Y) involves conducting a controlled experiment. In such an experiment, researchers intentionally change X while observing the resulting impact on Y, ensuring that other factors are held constant to confidently attribute the change in Y to the change in X.

The Scientific Standard for Establishing Causality: Controlled Experiments

The methodology of controlled economic experiments involves creating a setting that is as realistic as possible while precisely managing the conditions under which decisions are made. The core principle is to isolate the effect of a single factor by changing only that one condition and holding all other variables constant. This allows the experimenter to observe how decisions are affected by the specific change, thereby establishing a causal link.

Learn Before

Related