Add a lecture on correlation vs causation

[jstac]

Key ideas

• use relevant economic applications
• begin with graphical analysis
• add some kind of statistical analysis that confirms the ideas in the figures.

Seeking applications from Ippei Fujiwara (Keio)

[jstac]

Comments from Ippei:

When teaching first-year Econ students, I usually begin by presenting graphs with significant correlations. However, it is important to note that various stories can be told about the causation. Next, I introduce examples of natural experiments such as Regression Discontinuity Design, DID, Bunching Analysis, and Synthetic Control.

While the first part may be useful for motivation, I believe the second part aligns more closely with QuantEcon. I know of many interesting sites using R.

It is worth noting that while causal inference (the second part) allows us to identify causation only at the margin, structural models are necessary for more global results. I am currently thinking of a good example that would help undergraduate students easily understand the usefulness of the model.

[HumphreyYang]

A well-presented lecture series on causal inference using Python for future reference: https://matheusfacure.github.io/python-causality-handbook/landing-page.html

[HumphreyYang]

Suggestions for case studies from Ippei:

If you draw the scattered plots of policy interest rate and GDP (or inflation rate). We usually observe a positive correlation.
This is contrary to our understanding of monetary policy. Higher interest rates should lead to lower GDP. Yet, as you know, there is a reverse (two-way) causality: One is monetary policy to macroeconomy as mentioned above, and the other is macroeconomy to monetary policy, that is a policy reaction. So, in the scattered chart, we observe the dominance of the latter.

So, in order to understand the effectiveness of monetary policy, we need to find the case with surprise monetary policy change or make a model for counterfactual simulation.

Also, as a related topic, we know from the Fisher equation that higher inflation leads to higher nominal interest rate. However, again, the central bank increases the interest rate to lower inflation rate. To understand this, you need a model.