Steps of Econometric Model Formation

Econometric model formation is the systematic process of translating an economic theory into a quantitative (statistical) model, estimating it using data, testing its validity, and using it for economic analysis and decision-making.

Steps of Econometric Model Formation

1. Existing Theories (Previous Studies)

This is the starting point of econometric analysis.

• Economic theories and earlier empirical studies suggest relationships among variables.

• These theories guide the choice of dependent and independent variables.

• Example: Keynesian theory suggests consumption depends on income.

2. Formulation of an Estimable Model

At this stage, the theoretical relationship is converted into a mathematical and statistical form.

• Economic variables are expressed as equations.

• A stochastic error term is added to capture randomness and unobserved factors.

• Example:

  

  
  

3. Collection of Data

At this stage, researchers design the research methodology, that is how and what data to be collected. The relevant data required to estimate the model are gathered.

• Data may be time series, cross-sectional, or panel data.

• Sources include government publications, surveys, databases, etc.

• Data quality is crucial for reliable results.

  
  

4. Model Estimation

The unknown parameters of the model are estimated using statistical techniques.

• Common methods include Ordinary Least Squares (OLS), Maximum Likelihood, etc.

• This step provides numerical values for coefficients.

5. Is the Model Statistically Adequate?

The estimated model is tested for statistical validity.

• Diagnostic tests are conducted for:

  • Significance of coefficients

  • Goodness of fit

  • Autocorrelation, heteroscedasticity, multicollinearity

• The model must satisfy classical econometric assumptions.

5.1. If NO → Revise the Model

If the model fails statistical tests:

• Variables may be added or removed.

• The functional form may be changed.

• Data issues may be corrected.

• The process loops back to model formulation and estimation.

6. If YES → Interpret the Model

Once the model is statistically adequate:

• Estimated coefficients are interpreted economically.

• Direction and magnitude of relationships are analyzed.

• Results are compared with economic theory.

7. Use for Analysis

Finally, the model is applied for practical purposes such as:

• Forecasting

• Policy evaluation

• Hypothesis testing

• Decision-making by governments or firms

Note that Econometric model formation is iterative, not one-time. Therefore, Feedback and revision improve model reliability.

Case Study

Below is a clear, real-world, exam-friendly example for each step of Econometric Model Formation, using household consumption and income as the context (a standard example in economics).

1. Existing Theories (Previous Studies)

Example: Keynes’ Psychological Law of Consumption states that as income increases, consumption also increases, but by a smaller proportion. This theory suggests a relationship between household consumption (C) and disposable income (Y).

2. Formulation of an Estimable Model

Example: The theoretical idea is converted into a statistical equation:

Where:

• C = Household consumption expenditure

• Y = Disposable income

• u = Error term (tastes, habits, expectations, etc.)

This equation can be estimated using real data.

3. Collection of Data

For the given example, data on income and consumption are collected from: National Sample Survey (NSS), Household budget surveys, RBI or World Bank databases. Further, the data can be: Cross-sectional (many households in one year) or Time series (one economy over many years). What type of data will be collected will depend on the objective of the study.

4. Model Estimation

Using Ordinary Least Square (OLS) method, the proposed model can be estimated. Suppose, the estimated model takes the following form:

Interpretation:

• Autonomous consumption = ₹200

• Marginal propensity to consume (MPC) = 0.75

5. Is the Model Statistically Adequate?

To understand the statistical adequacy, the analyst needs to perform diagnostic tests like, t-test (to check the significance of the coefficient of income), R-square (to check the explanatory power of the model or goodness-of-fit) and diagnostic check for autocorrelation or heteroscedasticity.

5.1. If NO → Revise the Model

Alternative Example: If tests reveal problems (e.g., omitted variables), we need to revise the model by adding interest rate or wealth (W) or transforming variables (log form) or using better-quality data. Then re-estimate the model. A proposed alternative model is given below:

6. If YES → Interpret the Model

If we found the model is statistically acceptable, the next step is data interpretation. According to the given example, a ₹1 increase in income increases consumption by ₹0.75. This finding confirms Keynesian theory that income is a strong determinant of consumption. Remember, here Economic meaning is emphasised, not just Statistics.

7. Use for Analysis

The government can use the model to:

• Predict the impact of tax cuts on consumption

• Estimate multiplier effects

• Design fiscal policy to stimulate demand