Sunteți pe pagina 1din 10

20-Jul-10

Lecture objectives

SEEG 5013 Managerial Economics

o To examine some general difficulties


encountered in deriving the demand curve for
a product from market data.

Topic 3: Demand Estimation


Chapter 4: Demand Estimation

o To discuss some marketing research


approaches to demand estimation.
o To discuss regression analysis as the most
useful and common method of demand
estimation.

Dr. Hj. Mohd Razani Hj. Mohd Jali


FE 0.55 (Economics Building)
College of Arts and Sciences
razani@uum.edu.my
04-928 3524

The identification problem

The identification problem

Demand curve for commodity estimated


from market data on quantity purchased of
the commodity at various prices over time
(time series data) or for various consuming
units or markets at one point in time (crosssectional data).

What it is not:
Simply joining the price-quantity observations
on a graph does not generate the demand
curve for the commodity.
Reason:
Each price-quantity observations is given by
the intersection of a different (but observed)
demand and supply curve of the commodity.
3

The identification problem

The identification problem

Over time - Demand for the commodity shifts or


differs because of changes in tastes, incomes,
price of related commodities, etc.
Supply curve shifts or differs because of
changes in technology, factor prices and
weather conditions (for agricultural
commodities).

The equilibrium of different but unknown


demand and supply curves generates different
price-quantity points observed. Thus, by simply
joining the different price-quantity observations,
we do not generate demand curve for
commodity. It is not that simple. This is
referred to as the identification problem.

20-Jul-10

Marketing Research Approaches to


Demand Estimation
Consumer Surveys
data from survey questions

Observational Research
data from observed behavior

Consumer Clinics
data from laboratory experiments

Market Experiments
data from real market tests
7

Market research approaches to


demand estimation

Market research approaches to


demand estimation

Consumer surveys:
Involve questioning sample of consumers
about how they respond to changes in price of
commodity, income, price of related
commodities, advertising expenditures, credit
incentives and other determinants of demand.

Problem with consumer surveys:


Biased. Respondent cannot give accurate
answers to things that have not happen yet
(eg. Reaction to price change to commodity
consumption)
Can be very expensive.

Market research approaches to


demand estimation

10

Market research approaches to


demand estimation

Observational research:
Gathering of info on consumer preferences by
watching them buying and using products.
Using scanners at stores see how many
people normally buys the product or how
many people buys the product after the
commercial, etc.

11

Problems:
Observation alone not enough to understand
consumers response.
Cannot determine consumers demographic
characteristics age, sex, education, income,
etc.

12

20-Jul-10

Market research approaches to


demand estimation

Market research approaches to


demand estimation

Consumer clinics:
Laboratory experiments where participants
are given a sum of money and asked to spend
in a simulated store to see how they react to
changes in price, packaging, displays, price of
competing products, etc.
Participants can be selected to represent
socio-economic characteristics of the actual
market.

Problems:
Questionable results because its not real
market environment.
Cost of conducting experiments is high can
only use small sample. Cannot represent
market behavior.

13

Market research approaches to


demand estimation

14

Market research approaches to


demand estimation

Market experiments:
Conducted in actual marketplace.
Select several markets with similar socioeconomic characteristics and change product
price, packaging, promotion, etc. in different
stores; and record consumer responses
(purchases) in different markets.

Market experiments:
Can be conducted in large scale ensure
validity of the result.
Consumers not aware of the experiment,
would act naturally to changes.
Useful in process of introducing new product
where no data exist yet.

15

16

Market research approaches to


demand estimation

Market research approaches to


demand estimation

Problems:
Cost is high. Can be conducted on limited
scale or short period of time.
Cannot prevent extraneous occurrences bad
weather, may bias the result.
Competitors can sabotage the experiment by
changing price and other determinants under
their control.

Problems:
Competitors can also observe the experiment
and gain information that benefit them.
Firm might lose customer when they raise
price for their product in the experiment. The
customer may find another product that
he/she likes to replace the firms product.

17

18

20-Jul-10

Regression Analysis

Regression Analysis

Assume that a manager wants to determine


relationship between the firms advertising
expenditures and its sales revenue.
test the hypothesis that higher advertising
expenditures lead to higher sales and estimate
strength of the relationship.

Rate on advertising expenditures and sales


revenue:
Level of advertising expenditures (X)
independent variable.
Sales revenue (Y) dependent variable.

19

20

21

22

Regression Analysis

Regression Analysis
Year

10

44

40

11

42

12

46

11

48

12

52

13

54

13

58

14

56

10

15

60

Regression Analysis
Scatter diagram shows positive relationship
between level of firms advertising
expenditures and its sales revenue higher
advertisement associated with higher
revenue. The relationship is approximately
linear.

Scatter Diagram

23

24

20-Jul-10

Regression Analysis

Regression Analysis

Regression Line: Line of Best Fit


Regression Line: Minimizes the sum of the
squared vertical deviations (et) of each point
from the regression line.
Ordinary Least Squares (OLS) Method

To estimate the approximate linear


relationship between firms advertising
expenditures and sales revenue draw in line
that best fit between data points.

25

26

Regression Analysis
The slope of line will provide an estimate of
the increase in sales revenue that the firm can
expect with each $1 million increase in its
advertising expenditures. This gives rough
estimate of linear relationship between sales
revenue (Y) and advertising expenditures (X).
Y = a + bX

27

Regression Analysis

28

Regression Analysis

Y = a + bX
a is vertical intercept of linear relationship and
gives the value of Y when X = 0, while b is
slope of line and gives estimate of increase in
Y resulting from each unit increase in X.

Problem:
Different researcher/manager would fit a
somewhat different line to the same data
point and obtain different results.

29

30

20-Jul-10

Regression Analysis
Regression analysis is a statistical technique
for obtaining the line that best fit data points
according to an objective statistical criterion.
Thus, all researchers looking at the same data
would get exactly the same result.
Regression line is obtained by minimizing the
sum of the squared vertical deviations of each
points from the regression line called
ordinary least-squares method (OLS).
31

32

Exercise

Solution

You are a newly hired marketing trainee for a


corporation in the local community. Senior
management has wondered how other local
firms view your companys reputation. One
suggestion is to call a variety of managers from a
master contact list of firms that your firm has
done business with over the last 12 months and
ask them what they think of the company. How
reliable is this kind of survey method?

Using the firms own contact list is a very


limited, non-random sample of the population
of all local firms. The list suffers from sample
bias (firms already doing business are likely to
favorably view the firm) and probably response
bias (respondents might provide ultra-favorable
responses, what they believe the firm wants to
hear).

33

Exercise

34

Solution

What are the major advantages and drawbacks


of using controlled customer experiments to
determine demand?

An advantage is that experiments require


subjects to make actual decisions, rather than
indicating preferences and behavior. The results
are more likely to reflect true preferences. A
drawback is that subjects know they are part of
an experiment, and may not respond accurately.
Second, experiments tend to be of small scale
and of short duration, so that accuracy is
limited.
35

36

20-Jul-10

Exercise

Solution
The major advantage is that a firm can alter one or
more key decision variables in one market and
compare the outcome to another similar market in
which the variables did not change, or were
changed in a different manner. The method can
generate valuable information about pricing and
advertising policy. The main drawback is that all
other factors (including population size and
demographics, consumer incomes, tastes,
competitor prices) must be comparable. Of course,
this is not always possible. Controlled market
studies are also expensive to conduct.

What are the major advantages and drawbacks


of using controlled market studies to estimate
demand?

37

38

Exercise

Solution
Regression analysis uses uncontrolled data to
generate an equation that allows one to
measure the separate influences of multiple
explanatory variables (in the form of numerical
coefficients) on total demand. In addition, the
analysis provides statistics that measure the
accuracy (goodness of fit) of the equation.
Accordingly, the regression approach can
produce the same kinds of results as a carefully
controlled market study.

How can regression analysis use uncontrolled


data to estimate demand?

39

40

Simple Regression Analysis

Simple Regression Analysis


The Ordinary Least-Squares (OLS) Method
Objective: Determine the slope and intercept
that minimize the sum of the squared errors.

Ordinary Least-Squares (OLS)


Model:
Yt a bX t et

Yt a bX
t

n
t 1

)
e (Y Y ) (Y a bX

et Yt Yt

t 1

2
t

t 1

t 1

t 1 is the sum of all observations, from time


t=1 to t=n.
n

41

42

20-Jul-10

Ordinary Least Squares (OLS)

Ordinary Least Squares (OLS)

The estimated values of a and b are obtained by


minimizing the sum of the squared deviations.
The value of b is given by
n

(X
t 1

X )(Yt Y )

(X
t 1

X)

a Y bX

Estimation Example

Time

Xt

1
2
3
4
5
6
7
8
9
10

10
9
11
12
11
12
13
13
14
15
120

n 10

X
t 1

X 120
X t
12
10
t 1 n

Yt
44
40
42
46
48
52
54
58
56
60
500
n

Y 500

120

t 1

Y 500
Y t
50
10
t 1 n

Xt X

Yt Y

-2
-3
-1
0
-1
0
1
1
2
3

-6
-10
-8
-4
-2
2
4
8
6
10
n

(X
t 1

t 1

( X t X )2

12
30
8
0
2
0
4
8
12
30
106

4
9
1
0
1
0
1
1
4
9
30

X ) 2 30

106
b
3.533
30

X )(Yt Y ) 106

a 50 (3.533)(12) 7.60

(X

( X t X )(Yt Y )

43

Ordinary Least Squares (OLS)

Ordinary Least Squares (OLS)

Caution when using the regression line to


estimate sales revenue of the firm for advertising
expenditure very different from those used in the
estimation of the regression line itself.
Regression line should be used only to estimate
the sales revenue of a firm resulting from
advertising expenditures that were within the
range or that at least near the advertising values
that are used in the estimation of the regression
line.

Estimation Example

n 10
n

X
t 1

(X
t 1

t 1

Y 500
t 1

Y
t 1

X t 120

12
n
10

Yt 500

50
n 10

X ) 2 30

106
b
3.533
30

X )(Yt Y ) 106

a 50 (3.533)(12) 7.60

(X

120

t 1

46

Tests of significance of parameter


estimates

Ordinary Least Squares (OLS)


Regression analysis is based on a number of assumptions:
The error term
is normally distributed,
has zero expected value or mean,
has constant variance in each time period and for all values
of X,
its value in one time period is unrelated to its value in any
other period.

These assumptions are required so as to obtain


unbiased estimates of the slope coefficient and to be
able to utilize probability theory to test for the
reliability of estimates.
47

Estimate of slope coefficient from different


sample of adverts sales data would obtain
different result/estimate of slope coefficient.
The greater is the dispersion of the estimated
value of b, the smaller is the confidence that
we have in our single estimated value of the b
coefficient.

48

20-Jul-10

Tests of significance of parameter


estimates

Tests of Significance

To test the hypothesis that b is statistically


significant (adverts positively affects sales), we
need to calculate the standard error
(deviation) of b.
This can be done with regression analysis
using computer programs.

Standard Error of the Slope Estimate

(Y Y )
( n k ) ( X X )

e
(n k ) ( X

sb

2
t

X )2

49

Tests of Significance

Tests of Significance
Example Calculation

Example Calculation

Time

Xt

Yt

Yt

et Yt Yt

et2 (Yt Yt ) 2

( X t X )2

10

44

42.90

1.10

1.2100

40

39.37

0.63

0.3969

11

42

46.43

-4.43

19.6249

12

46

49.96

-3.96

15.6816

11

48

46.43

1.57

2.4649

12

52

49.96

2.04

4.1616

13

54

53.49

0.51

0.2601

13

58

53.49

4.51

20.3401

14

56

57.02

-1.02

1.0404

10

15

60

60.55

-0.55

0.3025

65.4830

30

e (Y Y )
t 1

2
t

t 1

65.4830

(X
t 1

(Y Y )
( n k ) ( X X )
2

X ) 30
2

sb

Tests of Significance
Calculation of the t Statistic

b 3.53

6.79
sb 0.52

Degrees of Freedom = (n-k) = (10-2) = 8


Critical Value at 5% level =2.306

t 1

2
t

t 1

(X
t 1

65.4830

0.52
(10 2)(30)

e (Y Y )
t

(Y Y )
( n k ) ( X X )
t

65.4830

X ) 2 30

sb

65.4830
0.52
(10 2)(30)

Tests of goodness of fit and correlation


Besides testing for statistical significance of particular
estimated parameter, we can also test for the overall
explanatory power of the entire regression.
This is accomplished by calculating the coefficient of
determination, denoted by R2 the coefficient of
determination.
Coefficient of determination is defined as the
proportion of the total variation or dispersion in the
dependent variable, that is explained by the variation
in the independent or explanatory variable in the
regression.
54

20-Jul-10

Tests of goodness of fit and correlation


R2 measures how much of the variation in the
firms sales is explained by the variation in its
advertising expenditures.
The closer the observed data points fall to the
regression line, the greater is the proportion
of the variation in the firms sales explained by
the variation in its adverts expenditures, and
the larger is the value of the coefficient of
determination, R2.

Questions or comments?

Reference:
Salvatore (2008), Ch. 4

55

10