The quest for causality in Economics, By João Pereira dos Santos

Psychologists often refer that persons who belong to the same group tend to have identical behavioral patterns. In addition, the causes for this to happen may be linked with endogenous or correlated effects. The former occur when a person acts in a specific way to conform to group norms and be accepted. The latter take place when individuals with similar personal characteristics choose to associate with each other. In this case, when they face analogous environments, they perform in a comparable manner.

Experienced econometricians know that it is sometimes very difficult to distinguish the above effects in the data. As Charles Manski wrote “it is similar to an inferential problem that occurs when one observes the almost simultaneous movements of a person and of his image in a mirror. Does the mirror image cause the person’s movements, does the image reflect the person’s movements, or do the person and image move together in response to a common external stimulus? Empirical observations alone cannot answer this question.” (Emphasis added)

The following step is to recognize why empirical strategies and the search for implication are relevant. A decent practical motivation is that this fact may involve conflicting public policy implications. For example, understanding how students cooperate in the classroom is key to assess the impact of reducing class sizes, curriculum programs and other educational procedures. Moreover, economists often consider that choice behavior reveals preferences. As a matter of fact, observation of a particular action only states that it is weakly preferred among several others.

Since statistical agencies may not have all the necessary resources to collect reliable data in practice that match what is required by numerous economic theories, important measurement errors may arise. Furthermore, there are some externalities difficult to determine. As Milton Friedman recognized, “the education of a child accrues not only to that child or to his parents but to the other members of the society. (…) Yet it is not feasible to identify the particular individuals benefited or the money value of the benefit (…). There is therefore a significant neighborhood effect”.

One way to overcome measurement errors is by using an instrumental variable (henceforth, IV) that is uncorrelated with them but, at the same time, is correlated with the measured variable. Although two-stage least squares method is consistent, it is not unbiased. Therefore, researchers should use large samples.

In order to avoid misleading interpretations, recent studies use IVs in estimates defining causal relationships and assign the variable of interest unsystematically. However, to force a randomly chosen group to, for example, attend one extra year of schooling is not always possible. Even more importantly, to find a proper IV may be a puzzling task (Bound et al., 1995). Researchers acknowledge that there is a difference between experiments they would like to make and those that legislation modifications and other “natural” factors allow them to do, using observational data. Some interesting studies in the literature are presented in the following table:

EXAMPLES OF STUDIES THAT USE IVs

 

Outcome Variable

Endogenous variable

Source of IVs

Reference

 

1. Randomized Experiments

 

Achievement test scores

Enrollment in private school

Randomly attributed school voucher

Howell et al., (2000)

 
 

Achievement test scores

Class size

Random assignment to a normal or small class

Krueger (1999)

 
 

Achievement test scores

Hours of study

Random mailing of preparation material

Powers and Swinton (1984)

 
 

2. Natural experiments

 

Earnings

Years of schooling

Region and time variation in school construction

Duflo (2001)

 
 

Earnings

Years of schooling

Distance to college

Card (1995)

 

Achievement test scores

Class size

Discontinuities in class size

Angrist and Lavy (1999)

 

To conclude, a misunderstanding of the implied directions of causality may lead political representatives to “bet on the wrong horse”. Hence, it is vital to have access to historical and micro data and perform sound analysis to understand where the efficient outcomes are.

About the author: João Pereira dos Santos is a student of the Masters in Economics at NovaSBE. He is currently working as a Research Assistant with Professor Maria Eugénia Mata, in the project The Portuguese Capital Market during the XXth Century. His areas of interest include International Political Economy, Macroeconomics, Public Economics and Economic History.

Advertisements

2 Comments

  1. It’s appropriate time to make some plans for the longer term and it is time to be happy.

    I have learn this put up and if I may just I desire to recommend you
    some interesting issues or advice. Perhaps you can write subsequent articles referring
    to this article. I desire to learn more issues about it!

  2. Wow. Fascinating viewpoint. I enjoyed how you protected this topic….
    A couple things I dont accept but hello… thats a different perspective.
    I am super-keen to examine the next post. Are you able to
    make the next one more in depth? Thanks 🙂

Leave a comment or suggestion

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s