Everything You Ever Wanted to Know About Building Your Econometrics Model

There is little doubt that every economic theory we studied is based on theoretical modeling first. It sprouted from an idea which later on was served to the world with improvements; assumptions, hypothesis tests and results came into prominence. These models are all the simplified versions of reality. Starting from the most basic law of demand or the complex IS-LM-AS-AD model or ever generating models of the future all of them are models depicting reality as best they can be.

As a beginner, students feel overwhelmed while building models for their thesis. We decode here the eight steps and the seven facts to keep in mind before building your econometrics model.

Step 1. Statement of theory of hypothesis

  1. Use an economic theory: consumption increases as income increases
  2. Read economics news articles, for fresh ideas. We’ll discuss the pitfalls of reading literature reviews before building up your model later in this article.

Step 2. Specification of mathematical model

Model = consumption = b0 + b1 income

Step 3. Specification of the statistical / eco model

Model = consumption = b0 + b1 income deterministic in nature. Ecotrix modifies it to deterministic and stochastic components u with well defined probabilistic  properties

Y = b0 + b1 income + u; u represents factors that affect cons but not taken into account explicitly 

Step 4. Obtaining the data

Collect data from secondary or primary sources.

Primary source data involves survey questionnaires while secondary data source involves data sourced from the collectors. Like Census data. Readily available.

Step 5. Estimation of the parameter of the model

Run regression analysis on the equation  to get the estimator’s values. Slope coefficient and intercept

Step 6. Hypothesis testing

Determining the slope or intercept is statistically significant . Use null and alternate

Step 7. Forecasting

Predict given the value of the independent variable the dependent variable value

Step 8. Using the model for control or policy purposes

7 facts to keep in mind before you build your model

  1. To Generate Idea

Best place to look for ideas is newspapers, non-economic magazines, real life conversations, your own life experiences and watching how others behave. For example watch the movie A Beautiful Mind to discover how the idea for Nash equilibrium in Game Theory came into existence.

  1. Explain it to your grandmother or in her absence a layman

If you can explain a concept to your grandmother successfully, and then you have expressed your idea well. Hold the discussion with a layman and ask if they find it interesting. Generate as much insights as possible from these discussions.

  1. Remember every economic model involves economic agents, their choices and objectives given the constraints. Observe how they behave in different circumstances, is any discrepancy in their behavior temporary or permanent? Ask questions to yourself. Write down the answers in a notebook. While reading new ideas crop up, questions arise. Collect them when they are fresh.  
  2. Keep it Simple, stupid: The model must be reduced to the form where no further simplification is possible just like the absolute advantage theory of 2 goods and 2 countries model.
  3. You can use Generalize to Specific or specific to generalize method. For example in case of specific to general you can use forward step regression. Relatively simple economic situation: wage = alpha + error

Add a next variable wage = alpha+ b1 education + error

If b1 statistically significant 

Wage = alpha + b1 edu + b2 exp + + e; keep adding extra variables checking if the model congruent with the reality, along with statistically significant variable coefficients.

Or you can use as many variables and reduce the model to specific ones. This is the method of modeling from generalization to specific.. This method is better than its counterpart which faces omitted variable bias.

  1. when to look for literature support
    1. After you have run regression, get results then you go for literature support. Consulting literature at the initial stage hampers one’s originality to a large extent. Also, literature is not fresh, newspaper, magazines contain fresh ideas.
  2. Present in Seminar and incorporate the suggestions and criticisms
    1. Present your paper in seminars to gather suggestions, criticisms and shortcomings. Incorporate those and now it’s time to serve to the world

Want to know about the 13 most common econometrics model used by researchers and students. Click here

Want to find more about the top 3 model specification tests in econometrics? Click here

Discover the commonly encountered problems in econometrics? Visit here

Curious about how to interpret regression coefficients? Read it here if still need help watch this video

Here is a reading for understanding the procedure for heteroscedasticity test

Leave a Comment

Your email address will not be published. Required fields are marked *

1 + sixteen =