What is Econometrics

1
Introduction
2
What is Econometrics

• Application of statistical methods to economics.
• It is distinguished from economic statistics
  (statistical data) by the unification of economic
  theory, mathematical tools, and statistical
  methodology.
• It is concerned with (1) estimating economic
  relationships (2) confronting economic theory
  with facts and testing hypotheses about economic
  behavior, and (3) forecasting economic variables .

3
Estimating Economic Relationships

• Examples include
• d/s of various products and services
• firms wishes to estimate the effect of
  advertising on sales and profits
• relate stock price to characteristics of the firm
• macro policy, federal, state, and local tax
  revenue forecasts

4
Testing Hypotheses

• Examples include
• Has an advertising campaign been successful in
  increasing sales?
• Is demand elastic or inelastic with respect to
  price-important for competition policy and tax
  incidence, among other things.
• Effectiveness of government policies on macro
  policy.
• Have criminal policies been effective in reducing
  crime?

5
Forecasting

• Examples include
• Firms forecast sales, profits, cots of
  production, inventory requirements
• Utilities project demand for energy. Sometimes,
  these forecasts arent very good, such as what is
  currently happening in California.
• Federal government projects revenues,
  expenditures, inflation, unemployment, and budget
  and trade deficits
• Municipalities forecast local growth.

6
Uncertainty in These Three Steps

• The reason is that we generally base these steps
  on sample data rather than a complete census.
• Therefore, estimated relationships are not
  precise.
• Conclusions from hypothesis tests may accept a
  false hypothesis or reject a true one.
• Forecasts are not on target.

7
CODING.XLS

• Represents responses from a questionnaire
  concerning the president's environmental
  policies.
• The data set includes data on 30 people who
  responded to the questionnaire.
• The data is organized in rows and columns.

8
(No Transcript)
9
Observations

• An observation is a member of the population or
  sample. Alternative terms for observations are
  cases and records.
• Each row corresponds to an observation. The
  number of observations vary widely from one data
  set to another, but they can all be put in this
  format.
• In this data set, each person represents an
  observation.

10
Variables

• Each column represents a variable. An alternative
  term for variable that is commonly used in
  database packages is field.
• In this data set, each piece of information about
  a person represents a variable. The six variables
  are persons age, gender, state of residence,
  number of children, annual salary and opinion of
  the presidents environmental policies.

11
Variables -- continued

• The number of variables can vary widely from one
  data set to another.
• It is customary to include a row that gives
  variable names.
• Variable names should obviously be meaningful -
  and no longer than necessary.

12
Type of Data

• There are several ways to categorize data.
• Numerical versus categorical
• Cross-sectional versus time series
• Using this example we can look at the various
  types of data.
• On the next slide is an alternate way to
  represent the data set.

13
(No Transcript)
14
Numerical versus Categorical

• The basic distinction between the two is whether
  you intend to do any arithmetic on the the data.
  It makes sense to do arithmetic on numerical
  data.
• Clearly, the Gender and State variables are
  categorical and the Children and Salary variables
  are numerical. Age and opinion variables are more
  difficult to categorize.
• Age is expressed numerically, and we might want
  to perform some arithmetic on age such as the
  average age of respondents. However, age could be
  treated as categorical.

15
Numerical versus Categorical -- continued

• The Opinion variable is expressed numerically on
  a 1-5 Likert scale. These numbers are only codes
  for the categories strongly disagree, disagree,
  neutral, agree, and strongly agree. It is not
  intended for arithmetic to be performed on these
  numbers in fact, it is not appropriate to do so.
  
• The Opinion variable is best treated as
  categorical.
• In the case of the Opinion variable there is a
  general ordering of categories that does not
  exist in the Gender and State variables.

16
Numerical versus Categorical -- continued

• We classify these types of variables as ordinal.
  If there is no natural ordering , as with the
  Gender and State variables, we classify the
  variables as nominal.
• Both ordinal and nominal variables are
  categorical.
• Categorical variable can be coded numerically or
  left in uncoded form. This option is largely a
  matter of taste.
• Coding a truly categorical variable doesnt make
  it numerical and open to arithmetic operations.

17
Numerical versus Categorical -- continued

• Some options for this example are to
• code Gender (1 for male and 2 for female)
• uncode Opinion variable
• categorize the Age variable as young (34 or
  younger), middle aged (from 35-59) and elderly
  (60 or older).
• The one performing the study often dictates if
  variables should be treated numerically or
  categorically there is no right or wrong way.

18
Numerical versus Categorical -- continued

• Numerical variables can be subdivided into two
  types - discrete and continuous.
• The basic distinction between the two is whether
  the data arises from counts or continuous
  measurements.
• The Children variable is clearly discrete whereas
  Salary is best treated as continuous.
• This distinction is sometimes important because
  it dictates the type of analysis that is most
  natural.

19
Cross-sectional versus Time Series

• Data can be categorized as cross-sectional or
  time series.The Opinion data is Example 2.1 is
  cross-sectional. A pollster sampled a cross
  section of people at one particular point in
  time.
• In contrast, time series data occurs when we
  track one or more variables through time. An
  example would be the series of daily closing
  values of the Dow Jones Index.
• Very different type of analysis are appropriate
  for cross-sectional and time series data.