PART 2 Demand analysis

Demand Theory , Estimation and

Forecasting
 1. Demand theory
What is Demand?
• The quantity of a good or service that customers are willing
and able to purchase during a specified period under a given
set of economic conditions
Two types of individual demand
1. Direct demand : it is derived from theory of consumer
behavior i.e directly related to bundles of consumption items for
utility maximization

2. Derived demand : is demand for factors of production such as

(labor, capital, land, entrepreneur)
 Law of Demand
• There is an inverse relationship
between the price of a good and the
quantity of the good demanded per time
period. Because of:

• Substitution Effect: as prices rise — or income

decreases — consumers will replace more expensive items with less
.
costly alternatives

• Income Effect: as income change, qt demand will be

changed
• Demand function: is the relation
between demand and factors
influencing its level

QdX = f(PX, I, PY, T)

 Individual Consumer’s Demand
QdX = f(PX, I, PY, T)
QdX = quantity demanded of commodity X
by an individual per time period
PX = price per unit of commodity X
I = consumer’s income
PY = price of related (substitute or
complementary) commodity
T = tastes of the consumer
 QdX = f(PX, I, PY, T)

QdX/PX < 0
QdX/I > 0 if a good is normal
QdX/I < 0 if a good is inferior
QdX/PY > 0 if X and Y are substitutes
QdX/PY < 0 if X and Y are complements
 Market Demand Curve
• Horizontal summation of demand curves of individual consumers.
But the following concepts may slightly effect shape of demand curve
and then elasticity.

• Bandwagon Effect: phenomena of mob motivation and mass

psychology towards one item i.e. fashion i.e demand curve will be
more elastic
• Snob Effect: It refers to the extent to which the demand for a
product is decreased owing to the fact that others are also
consuming the same product. demanding goods which is unique.
Demand curve less elastic

• Veblen effect: means the extent to which the demand for a product
is increased because it bears a higher rather than a lower price.
Demand curve will be less elastic. E.g giffen goods
Horizontal Summation: From
Individual to Market Demand
 Market Demand Function
QDX = f(PX, N, I, PY, T)
QDX = quantity demanded of commodity X
PX = price per unit of commodity X
N = number of consumers on the market
I = consumer income
PY = price of related (substitute or
complementary) commodity
T = consumer tastes
 Demand Faced by a Firm: is
determined by:
• Market Structure
– Monopoly
– Oligopoly
– Monopolistic Competition
– Perfect Competition
• Type of Good
– Durable Goods e.g automobile, refrigerators, washing
machine

– Nondurable Goods
– Producers’ Goods (demand for inputs i.e raw material,
capital, labor )- Derived Demand
 Linear Demand Function
QX = a0 + a1PX + a2N + a3I + a4PY + a5T

PX Intercept:
a0 + a2N + a3I + a4PY + a5T

Slope:
QX/PX = a1

QX
 Elasticity
Percentage change in a dependent
variable resulting from a 1 percent
change in an independent variable
 Price Elasticity of Demand

Q / Q Q P
Point Definition EP   
P / P P Q

P
Linear Function EP  a1 
Q
 Price Elasticity of Demand

Q2  Q1 P2  P1
Arc Definition EP  
P2  P1 Q2  Q1
Marginal Revenue and Price
Elasticity of Demand
Optimal Pricing Policy:

If price is factored out:

Thus the MR can be  1 

MR  P  1  
rewritten as  EP 
 Recall that profit maximization requires operating at
the activity level where marginal cost equals
marginal revenue. Most firms have extensive cost
information and can estimate marginal cost reasonably
well

The optimal profit max price

will be:

Q: With marginal costs of $25 and eP = –2, what is the

profit-maximizing price ???
Marginal Revenue and Price
Elasticity of Demand
PX
EP  1
EP  1

EP  1

QX
MRX
Ep,TR and MR
 Determinants of Price
Elasticity of Demand
Demand for a commodity will be more elastic if:
• It has many close substitutes
• More time is available to adjust to a price change
• Nature of the goods—luxury vs necessary
• Proportion of income spent on goods i.e if high
proportion of income spent ,then it is more elastic
 Determinants of Price
Elasticity of Demand
Demand for a commodity will be less
elastic if:
• It has few substitutes
• Less time is available to adjust to a
price change
Income Elasticity of Demand

Q / Q Q I
Point Definition EI   
I / I I Q

I
Linear Function EI  a3 
Q
Income Elasticity of Demand

Q2  Q1 I 2  I1
Arc Definition EI  
I 2  I1 Q2  Q1

Normal Good Inferior Good

EI  0 EI  0
Cross-Price Elasticity of Demand

QX / QX QX PY
Point Definition E XY   
PY / PY PY QX

PY
Linear Function E XY  a4 
QX
Cross-Price Elasticity of Demand

QX 2  QX 1 PY 2  PY 1
Arc Definition E XY  
PY 2  PY 1 QX 2  QX 1

Substitutes Complements
EXY  0 EXY  0
 Other Factors Related to
Demand Theory
• International Convergence of Tastes/preferences
– Globalization of Markets e.g coca cola is USA product
which become household item around the world.
– Influence of International Preference/as fashion on
Market Demand
• Growth of Electronic Commerce
– Online sales –increase coverage across the globe at very
shortest time with low transaction cost .eg. Gmarket, alibab,
amazon and etc.
– Online currencies i.e bitcoin, blockchain technology just using
binary bits, using RFID technologies such as VISA card, ATMA
card

– Customer Relationship Management: a company with

better CRM has more demand
 2. Estimation of Demand
• The demand curve for a commodity is generally
estimated from market data on the quantity
purchased of the commodity at various prices
over time or for various consuming units at one
point in time.

• The price-quantity relationship, after correction for

all the forces that shift the demand curve, gives
the true demand curve for the commodity.
 Two approach's for estimating demand
curve:

1. Marketing research approach: includes

a. Consumer survey: involves questioning a sample of consumers
about how they would respond to particular change in price of a
commodity, income, price of related goods, advertising
expenditure, credit incentives, and other determinants of demand.
This can be done at shopping centre.
E.g. How much your monthly beer consumption would change if the price of beer
increase by 10% per bottle, if your income rose by 10%

b. Consumer clinics: is laboratory experiments in which

participants are given some money and asked to spend in a
simulated store to see how they react to changes in price,
product packaging or etc.
c. Market experiment: is conducted in the actual
marketplace. E.g at different shopping mall
 Simply select several markets with similar socioeconomic
characteristic, and change the commodity price in one
market and change packaging services in some other
markets and promotion in the other market again and
record the different response of consumers in different
markets.
2. Using regression (Econometrics) analysis
Steps are:
a. model specification i.e. specify model to be estimated,
identify all variables that affect demand for the commodity
b. collecting the data on the variables i.e. time series or cross
sectional data across economic agents
c. specify the form of the demand equation i.e functional form
of demand function can be linear, or power function or log
linear fun.
d. evaluation of estimated result: using different criteria i.e
a. economic criteria: sign and size
b. Statistical criteria i.e goodness of fit and
standard error
c. econometrics criteria: check violations of assumptions
 Example for Econometrics
regression
• Suppose the quantity demand for a
teff(Y) is a function of its price(X1) and
consumer income(X2) and the data is
from 1997-2007 E.C as follows:
 Year Quantity Price of teff per Income of
• . demanded(Y) KG consumer
1991 72 10 2000
1992 81 9 2100
1993 90 10 2210
1994 99 9 2305
1995 108 8 2407
1996 126 7 2500
1997 117 7 2610
1998 117 9 2698
1999 135 6 2801
2000 135 6 2921
• Now if you estimate the demand for teff
using regression analysis:
• 𝑌𝑡 = 21.04 − 4.5𝑋1𝑡 + 0.05𝑋2𝑡
𝑅2 = 0.94
. reg quantitydemandedy priceofteffperkg incomeofconsumer

Source SS df MS Number of obs = 10

F( 2, 7) = 62.66
Model 4142.61903 2 2071.30951 Prob > F = 0.0000
Residual 231.380971 7 33.0544244 R-squared = 0.9471
Adj R-squared = 0.9320
Total 4374 9 486 Root MSE = 5.7493

quantityde~y Coef. Std. Err. t P>|t| [95% Conf. Interval]

priceoftef~g -4.522255 2.267505 -1.99 0.086 -9.884051 .8395417

incomeofco~r .050337 .0113115 4.45 0.003 .0235895 .0770845
_cons 21.04284 44.28109 0.48 0.649 -83.66529 125.751
 Exercise 2
Using elasticity in managerial decision
making:
1. Suppose the estimated market demand
for Jimma coffee brand(x) in Ethiopia
is as follows
𝑄𝑥 = 1.5 − 3.0𝑃𝑥 + 0.8𝐼 + 2.0𝑃𝑦 − 0.6𝑃𝑠 + 1.2𝐴
Where
 Qx-quantity sales of Jimma coffee/day
 Px-price of Jimma coffee brand/kg
 I-disposable income for consumers
 Py-price of competitive brand coffee
 Ps-price of sugar
 A-Advertisiment expenditure
Suppose this year, Px=2$, Py=1.80$, I=2.5$,
Ps=0.50$, A=1$.
• Then how much quantity demand of coffee will be
sold?
• Simply substitute:
𝑸𝒙 = 𝟏. 𝟓 − 𝟑(𝟐) + 𝟎. 𝟖(𝟐. 𝟓) + 𝟐(𝟏. 𝟖𝟎) − 𝟎. 𝟔(𝟎. 𝟓𝟎) +
𝟏. 𝟐(𝟏) = 𝟐 million tones of coffee

NB: using the above information, Jimma coffee produer

can identify the demand elasticity of its coffee to
each factors and predict for the next year if there
are changes:
 – EP=-3(2/2)= -3
– EI= 0.8(2.5/2)= 1
– Exy=2(1.8/2)=1.8
– EA=1.2(1/2)=0.6
• Suppose next year firm or coffee union intend to
increase its coffee price by 5%, advertisement
expenditure by 12%,price of sugar decreases by 8%,
and price of competitive firm increase by 7% and
disposable income increases by 4%, then find firms
or unions sales of coffee next year?
• New 𝑄’𝑥 = 𝑄𝑥 + 𝑄𝑥(Δ𝑃𝑥/𝑝𝑥). 𝐸𝑃 +
𝑄𝑥(Δ 𝐼/𝐼). 𝐸𝐼 + 𝑄𝑥(Δ 𝑃𝑦/𝑝𝑦). 𝐸𝑥𝑦 +
𝑄𝑥(Δ 𝑃𝑠/𝑝𝑠). 𝐸𝑥𝑠 + 𝑄𝑥(Δ𝐴/𝐴). 𝐸𝐴
• = 𝟐 + 𝟐(𝟓%)(−𝟑) + 𝟐(𝟒%)(𝟏) + 𝟐(𝟕%)(𝟏. 𝟖) + 𝟐(−𝟖%)(−. 𝟏𝟓) + 𝟐(𝟏𝟐%)(𝟎. 𝟔)
• =2.2 million tones of coffee
 Numerical estimation
• Suppose the demand function for the automobile industry is
𝑄 = 𝑎1𝑃 + 𝑎2𝑃𝐼 + 𝑎3𝐼 + 𝑎4𝑃𝑜𝑝 + 𝑎5𝑖 + 𝑎6𝐴

Where Q, is a linear function of

 the average price of new domestic cars (in $), P;
 the average price for new import cars (in $), PI;
 disposable income per household (in $), I;
 population (in millions), Pop;
 average interest rate on car loans (in percent), i; and
 industry advertising expenditures (in $ millions), A.
The terms a1, a2, . . ., a6 are called the parameters of the
demand function
The estimated result:
𝑄 = – 500𝑃 + 210𝑃𝑋 + 200𝐼 + 20,000𝑃𝑜𝑝 – 1,000,000𝑖 + 600𝐴
a. If income is 15,000 and Qd=1000, find point EI?
b. If domestic car price is 20,000 and Qd=1000, find EP?
c. With advertising exp of 20,000 and 10,000 unit sold, find point EA?
 Exercise for demand estimation
• Sales data for Branded Consumer Products

year-quarter
unit sales(Q)
price advrt expend
competitorsincome
price pop time variables
2000 193334 6.39 15827 6.92 33337 4116250 1
2001 170041 7.21 20819 4.84 33390 4140338 2
2002 247709 5.75 14062 5.28 33599 4218965 3
2003 183259 6.75 16973 6.17 33797 4226070 4
2004 282118 6.36 18815 6.36 33879 4278912 5
2005 203396 5.98 14176 4.88 34186 4359442 6
2006 167447 6.64 17030 5.22 35691 4363494 7
2007 361677 5.3 14456 5.8 35950 4380084 8
• Using the below equation estimate the
demand model.
• 𝑄𝑖𝑡 = 𝑏0 + 𝑏1𝑃𝑖𝑡 + 𝑏2𝐴𝑖𝑡 + 𝑏3𝑃𝑋𝑖𝑡 + 𝑏4𝑌𝑖𝑡 +
𝑏5𝑃𝑜𝑝𝑖𝑡 + 𝑏6𝑇𝑖𝑡 + 𝑢𝑖𝑡
a. Interpret the coefficients and
b. calculate the elasticity
 Assignment 2 (15%)
1. Demand Curves. ISHO-garment is contemplating a T-shirt advertising
promotion. Monthly sales data from T-shirt shops marketing indicate that
𝑄 = 1,500 – 200𝑃
• where Q is T-shirt sales and P is price.
a. How many T-shirts could ISHO-garment sell at $4.50 each?
b. What price would ISHO-garment have to charge to sell 900 T-shirts?
c. At what price would T-shirt sales equal zero?
d. How many T-shirts could be given away?
e. Calculate the point price elasticity of demand at a price of $5

2. Optimal Pricing. In an effort to reduce excess end-of-the-model-year inventory,

Harrison Ford offered a 2.5% discount off the average list price of Focus SE
sedans sold during the month of August. Customer response was enthusiastic, with
unit sales rising by 10% over the previous month’s level.
A. Calculate the point price elasticity of demand for Harrison Ford Focus SE
sedans.
B. Calculate the profit-maximizing price per unit if Harrison Ford has an average
wholesale cost of $10,000 and incurs marginal selling costs of $875 per unit.
 3. Demand forecasting
Q: Why demand forecasting? Are we sure how much qt to sell?

A firm must decide on

 How much of each product to produce?
 What price to charge
 How much to spend on advertising
 And plan for the future growth of demand for a firm
and all these are done through forecasting.

 The aim of forecasting: is to reduce the risk

or uncertainty that firms faces in short term
operational decision making and planning
for its long term growth.
 Demand Forecasting
Demand Forecasting is a systematic
• .
and scientific estimation of future
demand for a product. Simply,
estimating the sales proceeds or demand
for a product in the future is called as
demand forecasting.

Example: sales
forecast
Demand forecasting methods
• .
 A. Qualitative Forecasts(survey method)

• Survey Techniques
– Planned Plant and Equipment Spending
– Expected Sales and Inventory Changes
– Consumers’ Expenditure Plans
• Expert Opinion Polls
– Business Executives(managers view)
– Sales Force
– Consumer Intentions: users view
– panel consensus: Forecast method based on the
informed opinion of several individuals
– delphi method: Method that uses forecasts
derived from an independent analysis of expert
opinion
• This qualitative demand forecasts are mainly used for
shorter period forecasts because of inconsistency
and dynamics of consumer behavior
 B. Quantitative (statistical)
methods include:
1. Trend projection methods
2. Barometric methods
3. Econometrics methods
4. Input-output forecasting
 1. Trend projection methods

Trend analysis: is forecasting the future path of

• economic
. variables based on historical patterns of
time series

Forecasting by trend projection is predicated on

the assumption that historical relationships
will continue into the future.

There can be different types of trends based on

the type of time series economic data
 Source of variation in time
series data
a. Secular Trend: it refers to long-Run Increase or Decrease
in Data series. E.g series of sales exhibit rising trend b/s of
increase income, population growth
b. Cyclical Fluctuations: refers to long-Run Cycles of
Expansion and Contraction in most economic time series
data b/s of d/t reasons
c. Seasonal Variation: refer to regularly Occurring
Fluctuations in economic activity during each year
d. Irregular or Random Influences: e.g variation data from
wars, natural disaster, strikes and other unique events.
Unpredictable shocks to the economic system
A. linear trend analysis: Assumes
constant unit change over time
St = S0 + b t, where b = Growth per time period
Example .Microsoft sales data for the 1984–2001
period and the least squares regression method as
follows (t -statistics in parentheses):

it is possible to forecast firm sales for future periods. To do so, it is important to realize that in this
model, t= 1 for 1984, t= 2 for 1985, and so on. This means that t= 0 in the 1983 base period. To
forecast sales in any future period, simply subtract 1983 from the year in question to determine a
relevant value for t.

A sales forecast for Microsoft in the year 2010 is

t = 2010 – 1983 = 27
Sales 2010 = –$6,440.8 + $1,407.3(27) = $31,556 million
B. constant growth rate: Assumes constant percentage change
over time.
Constant Growth Rate
St = S0 (1 + g)t , where g = Growth rate
• Estimation of Growth Rate
lnSt = lnS0 + t ln(1 + g)
Using Microsoft sales data for the 1984–2001 period results in the linear
constant annual rate of growth regression model (t statistics in parentheses):

by transforming this estimated equation back to its original form:

*antilog=10 raised to the value

• .Thus, a constant annual rate of growth
model forecast for sales in 2010 is:

𝑡 = 2010 − 1983 = 27
𝑆2010 = $96.38(1.400)27
= $𝟖𝟓𝟎, 𝟎𝟒𝟗 𝒎𝒊𝒍𝒍𝒊𝒐𝒏
 2. Barometric methods
 Often, the barometric method of forecasting is used by the
meteorologists in weather forecasting. The weather
conditions are forecasted on the basis of the movement of
mercury in a barometer.
 Experts with intelligent information and good experience can
forecast.
 Based on this logic, economists use economic indicators as a
barometer to forecast the overall trend in the business
activities.

 It works based on an index of relevant economic

indicators and forecasting the future trends by
analyzing the movements in these indicators.
Relevant economic indicators are:
 Leading Indicators. E.g indicators which move up or down
ahead of some other series. e,.g, net business investment index,
corporate profits after tax, etc.
 Lagging Indicators: e.g. A series consisting of those indicators,
which after some time-lag follows the change. Some of the lagging
series are- outstanding loan, labor cost per unit production,
lending rate for short-term loans, etc.
 Coincident Indicators: indicators which move up and down
simultaneously with the general level of economic activities. The examples of
coincidental series – the rate of unemployment, sales recorded by
manufacturing, retail, and trading sectors, gross national product at constant
prices.
 Composite Index: grouping data index
 Diffusion Index: summarizing the common tendency of a group
of statistical series
The major limitations of this method are;

 First, Often the leading indicator of the variable to be

forecasted is difficult to find out or is not easily available.
 Secondly, the barometric technique can be used only for a
short-term forecasting.

 If a lot of individual lagging indicators, leading indicators or

coincidence indictors are there, we will use the composite
index to make decision
 3. Econometric Methods
It uses economic theory and mathematical and statistical
tools to forecast economic relations.

Advantages of econometrics forecasting over alternative

methods.
 force the forecaster to make explicit assumptions about the
linkages among the variables in the economic system
 It produces logical consistency in the forecast model and
increases reliability.
 the forecaster can compare forecasts with actual results
and use insights gained to improve the forecast model
 3. Econometrics forecasting:
Models
Single Equation Model of the
Demand For Cereal (Good X)
𝑄𝑋 = 𝑎0 + 𝑎1𝑃𝑋 + 𝑎2𝑌 + 𝑎3𝑁 + 𝑎4𝑃𝑆 + 𝑎5𝑃𝐶 + 𝑎6𝐴 + 𝑒

QX = Quantity of X PS = Price of other substitute crop

PX = Price of Good X PC = Price of complementary
goods
Y = Consumer Income
A = Advertising
N = Size of Population
e = Random Error
4. Input-Output Forecasting
Three-Sector Input-Output Flow Table

Producing Industry
Supplying Final
Industry A B C Demand Total
A 20 60 30 90 200
B 80 90 20 110 300
C 40 30 10 20 100
Value Added 60 120 40 220
Total 200 300 100 220
Input-Output Forecasting
a. Direct Requirements Matrix

Direct = Input Requirements

Requirements Column Total

Producing Industry
Supplying
Industry A B C
A 0.1 0.2 0.3
B 0.4 0.3 0.2
C 0.2 0.1 0.1
 Input-Output Forecasting
b. Total Requirement Matrix: is obtained by inverse
of Direct requirement matrix:
We use the Leontief matrix: (I-A), instead of original matrix:
1
(𝐼 − 𝐴)−1 = 𝐴𝑑𝑗(𝐼 − 𝐴)
(𝐼 − 𝐴)
Producing Industry
Supplying
Industry A B C
A 1.47 0.51 0.60
B 0.96 1.81 0.72
C 0.43 0.31 1.33
How to calculate inverse of matrix?
Reading assignment:

For example, 3x3 matrix

1. calculate matrix of minors
2. calculate matrix of cofactors(just apply sign to
the minor coefficients).
-(i+j)square
3. calculate adjoint matrix i.e transpose of
cofactor matrix
4. multiply 1/determinant by the cofactor matrix
to get inverse matrix

NB: for larger size of matrix we use computer

based “matrix calculator”.
Input-Output Forecasting

Total Final Total

Requirements Demand Demand
Matrix Vector Vector
1.47 0.51 0.60 90 200
0.96 1.81 0.72 110 = 300
0.43 0.31 1.33 20 100
• Using the total requirement matrix of the above
table, we can forecast the new total output of
industries A,B, and C.

• Fore example: if the final demand for the output of

industry A will increase from 90 to 100(from the
above table), using the total requirement matrix,
forecast the new total output of industries A,B, and
C?
Total requirement matrix final demand vector total demand vector

1.47 0.51 0.60 100 200

0.96     300 
 1.81 0.72 110   

0.43 0.31 1.33 
 
20   100 


To forecast the new total output of industries A,B and C?

1.47(100)+0.51(110)+0.60(20)= 215
0.96(100)+1.81(110)+ 0.72(20)= 310
0.43(100)+0.31(110)+1.33(20)=104
Shortcoming of input-output forecasting:
1. direct and total requirement matrix coefficient are assumed to be
constant and does not allow input substitution and commodity
substitution in consumption
2. Input -output tables are usually available with many years time
lags and need to be updated.
.

Thank you!
Any Q?