What is Econometrics?

ECON 3209 · Week 7, Lecture 1 · Kerala Agricultural University

Department of Development Economics, KAU

Autumn 2026

Learning Outcomes

Define econometrics and explain its role in economics.
Distinguish clearly between correlation and causation.
Understand why data are central to empirical economic analysis.
Recognise the three major types of economic data.
Connect abstract econometric ideas to Kerala agriculture and cooperative banking examples.

What is econometrics?

Econometrics combines economic theory, mathematics, and data.
It helps us test economic ideas using observed evidence.
It also helps us forecast, estimate relationships, and evaluate policies.
Without econometrics, many policy debates remain descriptive rather than analytical.
In short, econometrics turns economic questions into measurable empirical models.

A working definition

Econometrics is the application of statistical methods to economic data in order to estimate economic relationships, test theories, and evaluate policy effects.

The word contains both economics and metrics.
The goal is not only to describe data, but to make structured inferences.
Good econometrics requires both data quality and economic reasoning.

Why do data matter?

Economic theory gives us ideas, but data tell us whether those ideas fit reality.
Farmers, cooperatives, and markets differ across districts and over time.
Data help us measure these differences quantitatively.
Public policy decisions are stronger when backed by evidence.
Econometrics makes evidence systematic rather than anecdotal.

Correlation is not causation

Two variables may move together without one causing the other.
A third factor may influence both variables.
Reverse causality may also confuse interpretation.
Example: high fertilizer use and high rice yield may both reflect better irrigation.
Econometrics tries to isolate causal effects carefully.

Kerala example: yield and fertilizer

Suppose we observe that farms using more fertilizer also have higher rice yield.
Is fertilizer the cause, or do richer farms simply buy more inputs?
Rainfall, soil quality, irrigation, and farmer skill may all matter.
This is why economists build models instead of relying on raw correlation alone.
Causal questions require stronger design and assumptions.

A simple correlation example in Python

Three major data types

Cross-section data: many units observed at one point in time.
Time-series data: one unit observed over many periods.
Panel data: many units observed over many periods.
Each type supports different economic questions.
Next lecture will study these in detail.

Economic questions and variables

Econometrics starts from a clear question.
Then we define a dependent variable and one or more explanatory variables.
Example: does household income affect loan repayment in cooperative societies?
Example: does irrigation raise rice yield in Palakkad?
Variables must be measurable before they can be analysed.

Visualising a relationship

Typical steps in an econometric study

State an economic question.
Build a simple theoretical relationship.
Collect appropriate data.
Estimate a statistical model.
Interpret the results in economic language.
Check whether the conclusions are credible.

Small dataset structures in Python

What makes a good econometric question?

It is specific and measurable.
It names the outcome we care about.
It identifies the explanatory variable clearly.
It suggests a feasible data source.
It leaves room for testing alternative explanations.

From description to evidence

Descriptive statistics tell us what the data look like.
Econometrics goes further by estimating relationships systematically.
It does not replace economic judgment; it strengthens it.
In agriculture and banking, evidence-based decisions can improve outcomes.
This is why econometrics is a core skill for applied economics students.

“Econometrics is the better half of economics”

“Econometrics is the better half of economics. It is scientific, practical and useful.”

T. Krishna Kumar, Economic and Political Weekly, 7-13 February 2004, Vol. 39, No. 6.
The quote captures why this course combines theory, data, and policy analysis.
For KAU students, econometrics turns questions about crops, prices, credit, and rural livelihoods into evidence.

Indian Econometrics: A rich history

Kumar (2004) reminds us that econometrics has a deep Indian tradition.
The Indian Econometric Society (TIES) was founded in November 1970.
P C Mahalanobis linked statistics and planning, helped shape the Second Five-Year Plan, and built planning work through ISI.
C R Rao and G S Maddala are central figures in modern econometrics; Maddala also wrote some of the most widely used reference textbooks.
Indian economists such as Amartya Sen and Abhijit Banerjee also show how rigorous quantitative reasoning can influence policy and development debates.

Practical applications of econometrics

Econometricians estimate relationships behind balance of payments, industrial output, fiscal deficit, and exchange rates.
In a Kerala context, similar tools help study cooperative credit recovery, district-level crop output, and commodity price movements.
Following Engle and Granger (Nobel Prize, 2003), econometrics is used to analyse asset-price and stock-price volatility.
Kumar also highlights applications in patent infringement damages, pay equity cases, and car demand from product attributes.
The same logic can be used for Indian examples such as tractor demand, rubber price fluctuations, or salary equity across institutions.

Econometrics can be used strategically

Kumar notes that knowledge of econometrics also helps us detect deliberate misspecification.
Friedman’s permanent income hypothesis used omitted-variable logic to argue that simple Keynesian consumption equations overstated the multiplier.
The lesson is ethical as well as technical: if we omit relevant variables, coefficients may look persuasive but be misleading.
Good econometric training helps us both build better models and criticise weak ones.

Exercise

Create a small DataFrame with income_lakh and loan_repayment_rate for five borrowers. Compute the correlation matrix and write one sentence explaining why a positive correlation does not necessarily prove causation.

Summary

✅ Econometrics combines theory, mathematics, and data.
✅ It helps estimate relationships, test theories, and evaluate policy.
✅ Data are essential because economic claims must be checked against reality.
✅ Correlation is not the same as causation.
✅ Cross-section, time series, and panel data are the main data types.
✅ Clear questions and careful data choices are the starting point of good econometrics.

Next Lecture

We will study economic data types in more detail.
The focus will be cross-section, time series, and panel data with cooperative banking examples.