Econometrics Notes.

Econo = Economics

Metrics = Measurement

- N/B: Econometrics is a branch of economics that uses statistics, mathematics, and

economic theory to measure, test, and quantify economic relationships using real-
world data.
- Econometrics is the application of statistical methods to economic data in order to:

- What econometrics does

➢ Estimation: It estimates numerical relationships between variables
➢ Hypothesis testing: Econometrics tests whether economic theories hold in reality
➢ Forecasting: Using past data to predict future outcomes.
➢ Policy evaluation: It evaluates the impact of policies and interventions

Population vs Sample

Population:

- A population is the entire group of individuals, households, firms, or observations that

a researcher is interested in studying.
- Population parameters are fixed but usually unknown e.g (Population mean (μ),
Population variance (σ²), Population proportion (p)).

1
- Population mean: 𝜇 = 𝑁 ∑𝑁
𝑖=1 𝑋𝑖

Sample:

- A sample is a subset of the population, selected to represent the population.

- Sample statistics are calculated from data and are used to estimate population
parameters e.g Sample mean (X̄), Sample variance (s²), Sample proportion (p̂ )
 1
- Sample mean 𝑋ˉ = 𝑛 ∑𝑛𝑖=1 𝑋𝑖

Why Sampling??

➢ Cost efficiency: Studying the whole population is very expensive, while sampling
reduces data collection costs significantly.
➢ Time saving: Collecting data from a sample takes less time compared to surveying the
entire population.
➢ Feasibility: Some populations are too large or difficult to access, making full
population studies impractical.
➢ Adequate accuracy: A properly selected random sample can provide estimates that
closely approximate population values.
➢ Statistical inference: Sampling allows estimation of population parameters and
hypothesis testing using statistical methods.
VARIABLES

A variable is any measurable characteristic or quantity that can take on different values
(change) and is used in an economic model to explain or predict economic behaviour.

Types of Variables

▪ Dependent variable: the outcome variable that the econometric model seeks to explain
or predict. / the variable that is dependent on other variables

Example: Household food security, GDP growth.

▪ Independent/explanatory/ Regressor variable: Variables used to explain changes in

the dependent variable.

Example: Income, education, interest rate.

Y = 𝛼 + 𝛽*X + C

GDP = 𝛼 + 𝛽*

Continuous variable

A continuous variable can take any numerical value within a given range, including fractions
and decimals.

Example: Income, land size (hectares), GDP, height

Discrete variable

A discrete variable takes countable, distinct values, usually whole numbers.

Example: Number of household members, number of livestock.

Categorical variable

A categorical variable classifies observations into groups or categories with no numerical

meaning.

Example: Provinces, gender, education level, employment sector.

Dummy (binary) variable

A dummy variable is a numerical representation of a category, taking only values 0 or 1.

Example: Gender (1 = male, 0 = female), farming status.

Graph showing the relationships on the Slides……………

DATA

✓ Data are raw facts, figures, or observations collected about individuals, households,
firms, events, or phenomena, which can be measured, recorded, and analysed to
produce information and support decision-making.
✓ In economics and econometrics, data usually consist of numerical or categorical
observations on variables such as income, prices, employment, education, or output.

Types Of Data

1. Cross-sectional data
- Cross-sectional data consist of observations on many units (individuals, households,
firms, etc.) at a single point in time.

- Denoted by i
- Data collected at a single point in time on or more variables and provides a snapshot
of observation
Example: General Household Survey (GHS) 2016.

2. Time series data

- Time series data consist of observations on a single unit or aggregate measured over
several time periods.

- Data collect on a single unit over a period of time and is based on specific frequency
- Denoted by t
Example:

FarmPrice𝑡 = 𝛼 + 𝛽 PrimeRate𝑡 + 𝑢𝑡 ,

𝑡 = 1994–2008
3. Panel data (Longitudinal data)
- Panel data involve observations on the same units repeatedly over time, allowing for
analysis of both cross-sectional and time-series variation.

- Denoted by it

Example: National Income Dynamics Study (NIDS), which revisited the same approximately
7,500 households in 2008, 2010, 2012, 2014, and 2017.
STATISTICS PART

Mean

The mean / the average is the central value of a dataset.

Where

• Xi = each observation
• n = number of observations
• ∑ = sum of all values

Variance

Variance measures how spread out the data is from the mean.

Population Variance

Sample Variance
Standard Deviation

It is the square root of the mean

Correlation

Correlation measures the strength and direction of a linear relationship between two variables.

It usually takes the values between 1 and -1

r = 1 ., Perfect positive relationship

r = -1., perfect negative relationship

r = 0 ., no linear relationship

REGRESSION

- It is the relationship between a variables (dependant variable and one or more

independent variables).
- In simple terms, regression helps us understand:

How one variable affects another

By how much the dependent variable changes when an independent variable changes
 𝒀 = 𝜶 + 𝜷𝑿 + 𝒖

Where:

• 𝑌= dependent variable (outcome of interest)

• 𝑋= independent variable (explanatory variable)

• 𝛼= intercept

• 𝛽= slope coefficient

• 𝑢= error term / stochastic error term.

Formulas

Solving for the Intercept

𝑎 = 𝑦ˉ − 𝑏𝑥ˉ

Solving for the Slope

𝑁
∑𝑖=1( 𝑥𝑖 − 𝑥ˉ)(𝑦𝑖 − 𝑦ˉ)
𝑏= 𝑁
∑𝑖=1( 𝑥𝑖 − 𝑥ˉ)2
 Example in economics

𝐼𝑛𝑐𝑜𝑚𝑒𝑖 = 𝛼 + 𝛽 Education𝑖 + 𝑢𝑖
For i = 1 to N

𝐼𝑛𝑐𝑜𝑚𝑒𝑡 = 𝛼 + 𝛽 Education𝑡 + 𝑢𝑡
For t = 1 to N

𝐼𝑛𝑐𝑜𝑚𝑒𝑖𝑡 = 𝛼 + 𝛽 Education𝑖𝑡 + 𝑢𝑖𝑡

For it = 1 to N

Simple regression

- It is a regression with one dependent variable and one independent variables.

Wage = β0 + β1 Education + u

GDP = β0 - β1 Inflation + u

Multiple Regression

- It is a regression with one dependent variable and two or more independent variables.

Wage = β0 + β1Education + β2 Experience + β3Age + βGender + u

Why would you have Two or more independent variables?

- To control for other factors that also influence the dependent variable (to isolate the
ceteris paribus effect).
- To avoid omitted variable bias
- To improve the explanatory power and accuracy of the model
Population Regression (Line)

- A population regression (line) is the regression one would run if one had population
data.
- the ‘population regression line’ is the fitted line based on the population regression.
- One calculates the true intercept and slope directly from the population data, e.g. ‘α’
and ‘β’.

Sample Regression (Line)

- A sample regression is the regression one would run if one had only sample data (which
is usually the case)
- the ‘sample regression line’ is the fitted line based on the sample regression,
- In sample one estimates the intercept and slope, i.e. one finds ‘𝛼̂’ and ‘𝛽̂’.

Population and Sample

Stochastic Error term / Error Term / Residual

- the vertical gap between the value of Y and the predicted Y in terms of the population
regression
- It accounts for all variation on the dependant variable that are not included in the
explanatory variables.
 - It is a random variable in a regression model that captures the effects of all unobserved
factors, omitted variables, measurement errors, and random shocks that influence the
dependent variable but are not explicitly included as explanatory variables in the model.

Why there is an error term

In a regression model:

𝑌 = 𝛽0 + 𝛽1 𝑋 + 𝑢

The error term 𝑢exists and is included because:

1. Omitted variables:

Not all factors that affect 𝑌are included in the model. For example, wages depend on education
(𝑋) and experience, ability, and motivation. If we don’t include all of them, their effects go
into 𝑢.

2. Measurement errors:

Variables may not be measured perfectly. Any discrepancies between true values and observed
values are captured by 𝑢.

3. Random shocks:

Unpredictable events (like luck, sudden illness, or natural shocks) affect 𝑌.

4. Model simplification:

Economic models simplify reality; they cannot include every factor influencing 𝑌. The error
term captures the leftover unexplained variation.
In short: the error term accounts for all influences on 𝑌that are not included in the
explanatory variables.

T=54359+3.22L

^=12614+1.75L+1.23K

T^=6125+0.93L+0.06K+1.55W