TOC
Introduction
Brief indentification of Econometrics:
- Combine Statistical techniques with economic theory
- Estimating economic relationships
- testing economic theories
- evaluating and implementing government and business policy
Three Types of Econometrics Questions
1. Descriptive
Examples:
- What is the tax schedule with regard to income?
- What is the price level of consumer goods last quarter?
Two challenges for this kind of questions:
- Sampling
Since we typically do not observe the full population but rather a sample. We need to draw conclusion about the population based on the sample.
- Summary statistics
Some data are complicated, so we need to come up with a nice way to summarize them.
If we had data we would know the answer.
2. Forecasting
Examples:
- What will the GDP growth rate be next year?
- How much revenue will my firm earn next month?
Two challenges for this kind of questions:
- Under-fitting: models do not explain the current data well.
- Over-fitting: models explains the current data too well.
If we wait long enough and have data, we would know the answer.
3. Causal (or structural)
Examples:
- If the central bank lowers interest rates today, what will happen to inflation tomorrow?
- How much more money will you earn as a result of taking this course?
Correlation & Causation
Correltaion implies that how two random variables move together. We cannot decide whether two variables have causation directly. Causal interpretation can be given that is consistent with results derived from an appropriate statistical procedure. Having an economic model is often essential in establishing the causal interpretation.
Unless we can run the perfect experiment, we will never know the answer for sure.
The Structure of A Dataset
Terms for 2-D table
- Each column represents a variable
- Each row represents an observation
- Each cell represents a value
1. Cross-Sectional Data
- Each observation is uniquely determined by an unit (which can be an individual, a firm , a country, etc.)
- It consists of a sample of units taken at a given point in time.
- We typically assume that the sample is drawn from the underlying population randomly. While there maybe violation of random sampling.
2. Time Series Data
- Each observation is uniquely determined by time.
- A time series dataset consists of observations on a variable or several variables (of one unit) over time.
3. Pooled Cross Setions & Panel Data
Pooled Cross Sections:
- Pooled cross sections include cross-sectional data in multiple years.
- Different time, different units (not exactly the same).
Panel Data๏ผ
- Consists of a time series for each cross-sectional member in the data set.
- Different time, the same untis.
For Pooled Cross Sections & Panel Data, each observation is uniquey determined by the unit and the time.
ย
Loading Comments...