Introduction to Econometrics – Unit 1
Notes
Part A: Concise Notes
Definition of Econometrics
Econometrics is the application of statistical and mathematical methods to economic data to
give empirical content to economic theories and test hypotheses.
Scope of Econometrics
Includes testing economic theories, forecasting, policy evaluation, decision making, and
model building.
Methodology of Econometric Research
Formulate model → Specify econometric form → Collect data → Estimate parameters → Test
hypotheses → Forecast/policy analysis → Revise model.
Historical Origin of Regression
Introduced by Sir Francis Galton (1886) in context of heredity (heights). Modern meaning:
statistical tool for estimating relationships.
Statistical vs. Deterministic Relationship
Deterministic: exact (e.g., Area = πr²). Statistical: includes randomness (e.g., Consumption =
α + β Income + u).
Regression vs. Causation
Regression = association. Causation = direct cause-effect. Econometrics distinguishes using
theory/instruments.
Regression vs. Correlation
Correlation measures association (symmetric). Regression explains dependence
(directional).
Terminology and Notation
Dependent variable (Y), independent variables (X), parameters (β), error term (u), sample
size (n).
The Regression
Simple regression: Y = β0 + β1X + u. Multiple regression: Y = β0 + β1X1 + β2X2 + ... + βkXk +
u.
Nature and Sources of Data
Types: cross-sectional, time series, panel. Sources: government, surveys, experiments,
international organizations.
Part B: Detailed Notes
Definition of Econometrics
Econometrics combines economics, statistics, and mathematics to test theories and analyze
economic problems.
Example: Testing if consumption increases with income using regression analysis.
Scope of Econometrics
1. Testing theories (e.g., Phillips Curve).
2. Policy evaluation (fiscal, monetary).
3. Forecasting (GDP, inflation).
4. Model building.
5. Business decision making (demand, pricing).
Methodology of Econometric Research
Steps: (1) Statement of theory, (2) Specification of model, (3) Data collection, (4) Estimation
of parameters (OLS, MLE),
(5) Hypothesis testing (t-test, F-test), (6) Prediction/forecasting, (7) Model evaluation and
revision.
Historical Origin of Regression
Introduced by Francis Galton (1886) in heredity studies. Observed regression to the mean
in heights.
Now refers to estimation of statistical relationships between variables.
Statistical vs. Deterministic Relationship
Deterministic: exact (e.g., Area = πr²). Statistical: probabilistic, allows error term (e.g., Qd =
β0 + β1P + u).
Most economic relations are statistical due to human behavior and measurement errors.
Regression vs. Causation
Regression shows association, not causality. Example: Ice cream sales and drowning related
through temperature.
Causality tested using instrumental variables, experiments, Granger causality tests.
Regression vs. Correlation
Correlation = degree of linear association (symmetric).
Regression = explains dependence (directional). Regression of Y on X ≠ Regression of X on Y.
Terminology and Notation
Dependent variable (Y), independent variables (X), parameters (β0, β1, ...), error term (u),
sample size (n).
The Regression
Simple regression: Y = β0 + β1X + u. Multiple regression: Y = β0 + β1X1 + β2X2 + ... + βkXk +
u.
OLS estimation minimizes squared errors to estimate parameters.
Nature and Sources of Data
Types:
1. Cross-sectional (household survey at one point of time).
2. Time series (GDP from 1960–2025).
3. Panel data (income of households over years).
Sources: government agencies (CSO, RBI, NSSO, IMF, World Bank), surveys, experiments,
private databases, international sources.
