2 Introduction to the course

2.1 What is econometrics?

  1. Estimating economic relationships
    1. Demand curve \(\log(Q_{t})= \alpha_0 + \alpha_1 P_t + \epsilon_t\)
    2. Production function \(Y_{it}=A_{it}K_{it}^{\alpha}L_{it}^{\beta}\)
  2. Testing economic theory
    • Does adverse selection exists in insurance markets?
    • Are consumers rational?
  3. Determine the effect of a given intervention (causal inference)
    • What is the effect of increasing minimum wage on employment?
    • Do mergers increase the output price?
    • Does democracy cause economic growth? (a series of works by Acemoglu, Robinsohn, and their co-authors).
    • Effects of going to private colleges on your future earnings.
    • Note: Some questions may have underlying economic models, others may not.
  4. Describe the data (prediction/forecasting)
    • How does the distribution of wage look like?
    • Relationship between electricity consumption and temperature (possibly nonlinear).
    • Related to machine learning (ML).

2.2 Why do we need to learn computation

  1. Conduct statistical and empirical analysis using your own data set
    1. Construct the data set
    2. Describe the data
    3. Run regression or estimate an economic object
    4. Make tables and figures that show the results of your analysis.
  2. Verify the econometric theory through numerical simulations.
    • Ex. Asymptotic theory considers the case when the sample size is large enough (i.e., \(N \rightarrow \infty\))
      • Law of large numbers, central limit theorem
      • How well is the asymptotic approximation?
    • Monte Carlo simulations
  • We will learn both aspects in this course.

2.3 Why do we use R?

  • Many alternatives: Stata, Matlab, Python, etc…
  1. Free software!!
    • Stata and Matlab are expensive.
    • Though you can use Matlab through the campus license from this April.
  2. Good balance between flexibility in programming and easy-to-use for econometric analysis
    • Stata is easy to use for econometric analysis, but hard to write your own program.
    • Matlab is the opposite.
    • You can do everything with R, including data construction, regression analysis, and complicated structural estimation.
  3. Many users
    • Popular in engineering.
    • Many packages being developed (especially important for recently popular tools. )
  • Note: Python seems also good, though I have not used it before.