Sumários
Binary Dependent Variable Models
22 junho 2022, 12:00 • Hugo Castro Silva
Binary Dependent Variable Models
- Limited dependent variable models
- Examples of research questions with limited y
- Why not use linear regression?
- The linear probability model
- Alternatives to LPM: Logit and Probit
- The latent variable model
- Estimating logit and probit: maximum likelihood estimation
- Interpreting logit and probit: partial/marginal effects
- Partial effect at the average
- Average partial effect
- Goodness of fit in binary dependent variable models
- LPM, Logit, or Probit?
Instrumental Variables and Two Stage Least Squares
8 junho 2022, 12:00 • Hugo Castro Silva
Instrumental Variables and Two Stage Least Squares
1. Instrumental Variables
- Motivation
- IV Assumptions
- The IV Estimator
- IV Estimation of the Multiple Regression Model
2. Two Stage Least Squares
- Using multiple instruments
- IV/2SLS with poor instruments
- Testing for endogeneity
- Testing overidentification restrictions
Panel Data Analysis II —Fixed Effects, Random Effects, Correlated Random Effects
1 junho 2022, 12:00 • Hugo Castro Silva
Panel Data Analysis II — Fixed Effects, Random Effects, Correlated Random Effects
1. Fixed effects
- The within transformation.
- Time-demeaning variables.
- Estimating the within transformation with pooled OLS (POLS)
- Application (example 14.1)
- Effects that change over time, using FE
- Application (example 14.2)
- The dummy variable regression
- Fixed Effects vs First Differencing
- Fixed effects with unbalanced panels
2. Random effects
- Random effects: motivation
- Random effects: quasi-demeaned transformation
- Application (Example 14.4)
- RE vs POLS
- RE vs FE
3. The Correlated Random Effects approach
4. Applying data panel methods to other data structures
Panel Data Analysis I — First Differences (FD)
25 maio 2022, 12:00 • Hugo Castro Silva
Panel Data Analysis I — First Differences (FD)
What is Panel Data?
- Panel data in the long format
- Advantages and disadvantages of panel data
- The unobserved effects model
First Differences
- The first difference (FD) transformation
- FD interpretation
- FD limitations
- Policy analysis with FD
- Differencing and effects that change over time
- Application examples
- Serial correlation in panel data: what it is, what are its consequences
- Testing for serial correlation after a first differences regression
- Positive vs. negative serial correlation
- Correcting for serial correlation: clustered standard errors.
- FD limitations