Untitled Quiz

Introduction to Econometrics

• Econometrics is about using economic theory, data, and statistical tools to answer "how much" type questions.
• Its literal meaning is "measurement in economics".
• Econometrics uses theory and data from economics, business, and social sciences, along with statistical tools.

Types of Data

• Econometricians use three types of data:
  • Time series data
  • Cross-sectional data
  • Panel data (a mix of time series and cross-sectional data)
• Data can be quantitative (e.g., profit, price, cost, demand) or qualitative (e.g., motivation, satisfaction, service quality).

Continuous and Discrete Data

• Continuous data can take on any value (e.g., profit).
• Discrete data takes on specific values, usually integers (e.g., number of employees, quantity sold).

Ratio, Interval, Ordinal, and Nominal Numbers

• Numbers can be classified as cardinal, ordinal, or nominal.
• Ratio numbers have numerical values with meaning and an absolute zero point (equal distance between values).
• Interval numbers have meaningful order but no absolute zero (equal intervals between values).
• Ordinal numbers show order or position, but not meaningful differences between values.
• Nominal numbers have no natural order.

Examples of Econometric Problems

• Testing if financial markets are informationally efficient (weak form).
• Evaluating if CAPM or APT are better models for predicting returns on risky assets.
• Measuring and forecasting bond return volatility.
• Determining the factors that affect bond credit ratings.
• Modeling long-term price and exchange rate relationships.
• Finding the optimal hedge ratio for oil.
• Testing technical trading rules.
• Examining if earnings or dividend announcements affect stock prices.
• Evaluating if spot or futures markets react more quickly to news.
• Forecasting the correlation between stock indices of different countries.

Steps in Econometric Modeling

• Start with existing economic theory or evidence.
• Develop an estimable theoretical model.
• Collect data.
• Estimate the model.
• Evaluate the model's statistical adequacy.
  • If not adequate, reformulate the model and repeat the process.
  • If adequate, interpret the model.

Bayesian vs. Classical Statistics

• Classical statistics involves postulating a theory, building a model, and testing it with data.
• Bayesian statistics involves developing theory and model together, using prior beliefs (probabilities) combined with data to form posterior probabilities.

Considerations When Reading Econometric Papers

• Does the paper develop a theoretical model or just use techniques for data mining?
• Is the data quality good and from a reliable source? Is the sample size large enough?
• Were the techniques used validly? Were appropriate diagnostic tests conducted for assumptions?
• Are the conclusions sensible and justified by the results? Is the importance of the results overstated?

Additional Points in Applied Econometrics

• Use common sense and accounting/finance theory. Theory isn't just for model development, also crucial for interpretation and identifying testable predictions.
• Understand the real-life context of the system (history, institutions, peculiarities).
• How were the data gathered?
• Keep models sensibly simple. Use appropriate sophisticated techniques, not just because they are novel.
• Carefully check the estimation results (signs, significance, magnitudes). Ensure results are consistent with theory.
• Evaluate the sensitivity of results to factors such as sample period, explanatory variables.
• Assess the differences from using robust estimators.