Economic Data Types

ECON 3209 · Week 7, Lecture 2 · Kerala Agricultural University

Department of Development Economics, KAU

Autumn 2026

Learning Outcomes

  1. Define cross-section, time-series, and panel data clearly.
  2. Identify suitable examples from cooperative banking and agriculture.
  3. Recognise how dataset structure affects econometric questions.
  4. Organise each data type correctly in pandas.
  5. Compare the strengths and limitations of the three data forms.

Why data structure matters

  • The type of data determines what variation we can study.
  • Some questions need differences across units.
  • Some questions need changes over time.
  • Some questions need both dimensions together.
  • Econometric method follows data structure.

Cross-section data

  • Cross-section data observe many units at a single point in time.
  • Units may be households, farms, banks, districts, or firms.
  • Example: 50 cooperative societies in Kerala in March 2026.
  • These data are useful for studying differences across units.
  • They do not directly show change over time.

Cross-section example in pandas

Time-series data

  • Time-series data observe one unit repeatedly across time.
  • The time unit may be daily, monthly, quarterly, or yearly.
  • Example: monthly rubber prices in Kerala from 2010 to 2024.
  • These data are useful for trends, cycles, and forecasting.
  • Time order matters strongly in time-series analysis.

Time-series example in pandas

Visualising a time series

Panel data

  • Panel data combine cross-section and time-series dimensions.
  • We observe many units over multiple time periods.
  • Example: district agricultural productivity from 2018 to 2024.
  • Panel data allow us to study both differences across districts and changes over time.
  • They are especially useful for policy evaluation.

Panel data example in pandas

Comparing the three data types

  • Cross-section answers: who differs from whom?
  • Time series answers: what changed over time?
  • Panel data answers: who changed, and when?
  • Panel data often give richer evidence, but they are harder to manage.
  • The data type should match the research question.

Useful panel operations in pandas

Cooperative banking examples

  • Cross-section: loan recovery rates across 50 societies in one year.
  • Time series: monthly deposits in one district cooperative bank.
  • Panel: branch-level recovery rates from 2018 to 2025.
  • Different structures reveal different policy insights.
  • Data type affects both storage and econometric technique.

Common data management issues

  • Time variables must be stored consistently.
  • Unit identifiers should be unique and stable.
  • Panel data often contain missing years for some units.
  • Sorting by unit and time is essential.
  • Clear identifiers reduce errors in merging and analysis.

Choosing the right data for a question

  • If you want district comparison at one time, use cross-section.
  • If you want trend or forecast, use time series.
  • If you want policy impact across districts and years, use panel data.
  • Better data structure leads to better identification.
  • Data design is part of econometric thinking.

Exercise

Create one small cross-section DataFrame and one small panel DataFrame for cooperative societies. Then compute the mean recovery rate for the cross-section and a district-wise average for the panel.

Summary

  • ✅ Cross-section data observe many units at one point in time.
  • ✅ Time-series data observe one unit over multiple periods.
  • ✅ Panel data observe many units over many periods.
  • ✅ Data structure determines what type of variation we can analyse.
  • ✅ Cooperative banking and agricultural datasets often appear in all three forms.
  • ✅ Good econometrics begins with choosing the right data format.

Next Lecture

  • We will derive simple OLS step by step.
  • The lecture will move from the population regression function to the sample regression line and the formulas for \(\hat{\beta}_0\) and \(\hat{\beta}_1\).