Introduction

Part I

Introduction

Seppo Pynnönen Econometrics I

Introduction

Econometrics

1 Introduction
Econometrics
Steps in Econometric Analysis
Types of Econometric Data
Causality, Ceteris Paribus
Key Terms

Seppo Pynnönen Econometrics I

Introduction

Econometrics

Econometrics is a discipline of statistics

specialized for using and developing mathematical and
statistical tools for
empirical estimation of economic relationships
testing economic theories
making economic predictions
evaluating government and business policy.
Data: Nonexperimental (observational)
Major tool: Regression analysis (in wide sense)

Seppo Pynnönen Econometrics I

Introduction

Econometrics

Why study econometrics?

Important to be able to apply economic theory to real world
data.
Theory may be ambiguous as to the effect of some policy
change, and in any case theory rarely tells us how large the
effect might be.
Forecasting economic variables (inflation, interest rates,
housing starts, and so on).

(Source: Scientific Word slides by Wooldridge, Accompanion to

Introductory Econometrics)

Seppo Pynnönen Econometrics I

Introduction

Steps in Econometric Analysis

1 Introduction
Econometrics
Steps in Econometric Analysis
Types of Econometric Data
Causality, Ceteris Paribus
Key Terms

Seppo Pynnönen Econometrics I

Introduction

Steps in Econometric Analysis

(a) Economic model

For example, consumption function

C = f (Y , W ), (1)

where C is consumption, Y is income, W is wealth, and f is

some function (typically unknown).
Of course there are other factors affecting consumption, but
equation (1) captures the essence of the problem.

Seppo Pynnönen Econometrics I

Introduction

Steps in Econometric Analysis

(b) Econometric model

Usually the economic model does not specify exactly the functional
form. In addition the economic model is assumed to be exact in the
simplified world satisfying the simplifying assumptions.
Consumption and Income Consumption and Wealth

● ●
● ●

● ● ● ●
● ● ● ● ●
●
● ● ●● ● ● ●● ● ●
●● ● ●
● ● ● ● ● ●
●● ●●●
● ● ● ● ● ● ●
●● ● ● ●●
●●
● ●● ● ●●● ●
● ● ● ● ●●●
● ●
●
●
● ● ●● ● ● ●
●●
●
●
●
●●
●
● ●
● ●● ● ●●
●● ●●
● ● ● ●●●●● ●
●● ●
●● ●●●● ●● ● ● ● ● ●●●● ●●●●● ● ●
● ●● ● ● ●●
●
● ● ●
●
●●
● ● ●● ● ●
● ●●
● ●●●●
●● ●●●● ● ● ●
●●
●
●●●●●
●●●●●● ●● ●
●●● ● ● ●
● ● ● ● ●● ● ● ● ●
● ●●●● ●●
●
●● ●
● ● ● ● ● ● ●●●●
●●●● ● ●
● ●● ● ● ● ●●● ● ●
● ●● ●
●● ●
●●
●● ●●● ● ●●
● ●● ● ●●
● ●
●
●●●●
●
●●●
●●
● ●
● ●
Consumption

● ●

Consumption
●●●●●
●●● ● ● ●
●●●●●
●●
● ●
●●● ●
●
●
● ● ● ●●●● ● ● ●●●
●●●
●
●
● ●
●● ●●
●●●●● ● ●
●● ● ● ●●●●
● ● ●●
●
● ●● ●● ●
● ● ● ●●●
● ●●●●
● ● ●
●
● ● ●● ● ● ● ●●
● ●
● ●
● ●
● ●● ●● ●● ●
● ● ● ●
●●●
●●
●●● ● ●
●●
● ●● ● ● ●●
●●
●
●●
● ●●●●
● ●● ●
●
●●●●●
●●● ●
●●●
●
●
●
● ● ●
● ●●
●
●
●
●
●
●
●●
● ●
●● ●●● ●
● ●● ●●●●
●
● ●●● ●
● ●●● ● ●
●
●●
●
●●●
●
●● ●●●
●●
●●
● ●
● ● ●● ● ● ●●
●●
●
●
●●●
●●●
●
●
●●
●●● ●
● ●
●●
●●●
● ● ● ●
●●
● ●●●●●
●
●●● ● ●● ●
●● ● ● ● ● ●
● ● ● ●
●
●
●●●●
●● ●
● ●
● ● ● ● ● ●
●●
● ●
●●●
●
●● ●●● ●●
●
● ●●
●●●● ● ●● ●● ● ●●
●●
● ●
●● ● ●
●●●
●●
●
●
●
●
●
●
●
●
●
●
●
●
● ● ●●
●●
●
●
●
●
●
●●
●
●●● ●●●
●● ●
●● ●
●
●● ●●
●● ●● ● ●
● ●●●●
●●●
●
● ●●
●
● ●●
●
●
●
●●● ●
●
●
● ●●
●
●● ●
●
●●
● ●
●
●
●●●●
●● ●●●● ● ● ●
● ●
●
●●
●●●● ●● ●
●●●
●● ●
●
● ● ● ● ● ●
●●
●●●
●● ● ●● ●
● ●● ●
●●●●
● ●●●
●●● ●● ●
●
●●●
●
●●
●
● ●●
● ●
●●● ●
●●
●
●
●●
●●●●
●
●●●
●
●●
● ●
● ●
●
●● ● ●● ● ● ● ● ●●
● ● ●●
● ●
●●
●
●● ●●●● ● ● ● ●●
●
● ●
● ● ● ● ●● ● ● ●
●●
●
●
●
●
●
●
●● ●●
● ●●●●
● ●
●
●●
●
●●●
●
●
●
●
●●●
●
●●
●
● ●
●● ● ● ● ● ●●
●
●
● ●
●● ●●● ● ●● ●● ● ● ●
●
● ● ● ● ● ● ●
● ● ●●
● ●
● ●
● ●

● ●

Income Wealth

Seppo Pynnönen Econometrics I

Introduction

Steps in Econometric Analysis

The task of econometrics is to turn the economic model to an

operational one. Usually this amounts to a linear approximation,

C = β0 + β1 Y + β2 W + u, (2)

where β0 , β1 , and β2 are parameters of the model, to be estimated

from the data, and u is (unobservable) random error or disturbance
term, which determines the stochastic properties of the model.

Seppo Pynnönen Econometrics I

Introduction

Steps in Econometric Analysis

Another popular specification in economics is the log-log model.

log(C ) = β0 + β1 log(Y ) + β2 log(W ) + u. (3)

where log is the natural logarithm. In (3) parameters β1 and β2

are elasticities. Coefficient β1 is the income elasticity of
consumption and β2 is the wealth elasticity of consumption. (Can
you prove this?)
Remark 1.1: Although the same symbols are used, parameters β0 , β1 , and
β2 in (2) and (3) are different. The same is true with the error term u.
Remark 1.2: Model (3) can be again consideed technically in the
framework of linear models by defining c = log(C ), y = log(Y ), and
w = log(W ) so that (3) becomes

c = β0 + β1 y + β2 w + u,

which is again a linear model.

Seppo Pynnönen Econometrics I
Introduction

Steps in Econometric Analysis

A third possibility could be

log(C ) = β0 + β1 Y + β2 W + u. (4)

Can you figure out how e.g. β1 can be interpreted here?

Seppo Pynnönen Econometrics I

Introduction

Types of Econometric Data

1 Introduction
Econometrics
Steps in Econometric Analysis
Types of Econometric Data
Causality, Ceteris Paribus
Key Terms

Seppo Pynnönen Econometrics I

Introduction

Types of Econometric Data

(a) Cross-sectional
Data collected at given point of time. E.g. a sample of
households or firms, from each of which are a number of
variables like turnover, operating margin, market value of
shares, etc., are measured.
From econometric point of view it is important that the
observations consist a random sample from the underlying
population.

Seppo Pynnönen Econometrics I

Introduction

Types of Econometric Data

Example 1
Cross-sectional data on wages and other individual characteristics
(Wooldridge; subset of wage1.raw).

obs wage educ exper tenure female married

1 3.10 11 2 0 1 0
2 3.24 12 22 2 1 1
3 3.00 11 2 0 0 0
4 6.00 8 44 28 0 1
5 5.30 12 7 2 0 1
6 8.75 16 9 8 0 1
. . . . . . .
. . . . . . .
. . . . . . .
525 11.56 16 5 1 0 1
526 3.50 14 5 4 1 0

Seppo Pynnönen Econometrics I

Introduction

Types of Econometric Data

(b) Time Series Data

A time series consist of observations on a variable(s) over
time. Typical examples are daily share prices, interest rates,
CPI values.
An important additional feature over cross-sectional data is
the ordering of the observations, which may convey important
information.
Another additional feature is data frequency which may
require special attention.

Seppo Pynnönen Econometrics I

Introduction

Types of Econometric Data

Example 2
Time data on minimum wage and other characteristics for Puerto Rico
(Wooldridge; prminwage.raw)
year avgmin avgwage avgcov prunemp prgnp
1950 0.198 0.398 0.201 15.4 878.7
1951 0.209 0.410 0.207 16.0 925.0
1952 0.225 0.421 0.226 14.8 1015.9
. . . . . .
. . . . . .
. . . . . .
1986 3.350 4.725 0.581 18.9 4281.6
1987 3.350 4.879 0.582 16.8 4496.7

year = years 1950-1987;

avgmin = weighted avg min wage, 44 indust
avgwage = wghted avg hourly wage, 44 indust
avgcov = wghted avg coverage, 8 industries
prunemp = PR unemployment rate
prgnp = PR GNP

Seppo Pynnönen Econometrics I

Introduction

Types of Econometric Data

(c) Pooled Cross-sections

Both time series and cross-section features.
An example is a data set where a number of firms are
randomly selected, say in 1990, and another sample is selected
in 2000. (I.e., data consist of two different random samples.)
If in both samples the same features are measured, combining
both years forms a pooled cross-section data set.
Pooled cross-section data is analyzed much the same way as
usual cross-section data.
However, many times it is important to pay special attention
to the fact that there are 10 years in between.
Usually the interest is whether there are some important
changes between the time points. Statistical tools are usually
the same as those used for analysis of differences between two
independently sampled populations.
Seppo Pynnönen Econometrics I
Introduction

Types of Econometric Data

(d) Panel Data

Panel data (longitudinal data) consists of time series for each
cross-sectional member (i.e., same individuals) in the data set.
That is one has series of history from each individual.

Seppo Pynnönen Econometrics I

Introduction

Types of Econometric Data

Example 3
Firm panel data on n = 157 firms for three years per firm (Wooldridge;
jtrain.raw.
fcode year employ sales avgsal
410032 1987 100 47000000 35000
410032 1988 131 43000000 37000
410032 1989 123 49000000 39000
410440 1987 12 1560000 10500
410440 1988 13 1970000 11000
410440 1989 14 2350000 11500
.
.
419483 1987 133 11000000 13957
419483 1988 108 11500000 14810
419483 1989 129 12000000 14227
419486 1987 80 7000000 16000
419486 1988 90 8500000 17000
419486 1989 100 9900000 18000

Seppo Pynnönen Econometrics I

Introduction

Causality, Ceteris Paribus

1 Introduction
Econometrics
Steps in Econometric Analysis
Types of Econometric Data
Causality, Ceteris Paribus
Key Terms

Seppo Pynnönen Econometrics I

Introduction

Causality, Ceteris Paribus

Causality: Cause and effect x → y

Ceteris Paribus: ”Holding other relevant factors fixed”.

Seppo Pynnönen Econometrics I

Introduction

Key Terms

1 Introduction
Econometrics
Steps in Econometric Analysis
Types of Econometric Data
Causality, Ceteris Paribus
Key Terms

Seppo Pynnönen Econometrics I

Introduction

Key Terms

Data
Experimental data
Nonexperimental, observational
Random sampling
Cross-sectional data
Time series data
Data frequency
Pooled cross section
Panel data
Model
Economic model
Causal effects
Ceteris paribus
Econometric model

Seppo Pynnönen Econometrics I