ESTIMATION AND

FORECASTING
TECHNIQUES
O K TO FA Y U D H A S U D R A J A D, P H D
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
DEMAND ESTIMATION:
MARKETING RESEARCH
APPROACHES
• Consumer Surveys
• Observational Research
• Consumer Clinics
• Market Experiments
• Virtual Shopping
• Virtual Management
STEPS IN DEMAND ESTIMATION

• Model Specification: Identify Variables

• Collect Data
• Specify Functional Form
• Estimate Function
• Test the Results
FUNCTIONAL FORM
SPECIFICATIONS
Linear Function:

QX = a0 + a1 PX + a2 I + a3 N + a4 PY + +e

Power Function: Estimation Format:

QX = a ( PXb1 )( PYb2 ) ln QX = ln a + b1 ln PX + b2 ln PY
THE IDENTIFICATION PROBLEM
REGRESSION ANALYSIS
Year X Y
1 10 44
Scatter Diagram
2 9 40
3 11 42
4 12 46
5 11 48
6 12 52
7 13 54
8 13 58
9 14 56
10 15 60
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

REGRESSION ANALYSIS
ORDINARY LEAST SQUARES
(OLS)
Model: Yt = a + bX t + et

ˆ
Yˆt = aˆ + bX t

et = Yt − Yˆt
ORDINARY LEAST SQUARES
(OLS)
Objective: Determine the slope and
intercept that minimize the sum of
the squared errors.

n n n

t  t t  t
e 2

t =1
= (Y −
t =1
Yˆ ) 2
= (Y
t =1
− ˆ
a − ˆ )2
bX t
ORDINARY LEAST SQUARES
(OLS)
Estimation Procedure
n

(X t − X )(Yt − Y )
bˆ = t =1 ˆ
â = Y − bX
n

(X
t =1
t − X) 2
ORDINARY LEAST SQUARES
(OLS)
Estimation Example
Time Xt Yt Xt − X Yt − Y ( X t − X )(Yt − Y ) ( X t − X )2
1 10 44 -2 -6 12 4
2 9 40 -3 -10 30 9
3 11 42 -1 -8 8 1
4 12 46 0 -4 0 0
5 11 48 -1 -2 2 1
6 12 52 0 2 0 0
7 13 54 1 4 4 1
8 13 58 1 8 8 1
9 14 56 2 6 12 4
10 15 60 3 10 30 9
120 500 106 30
n n n
n = 10  X t = 120
t =1
 Yt = 500
t =1
(X
t =1
t − X ) 2 = 30 bˆ =
106
30
= 3.533

n n n
X t 120 Yt 500
X = = = 12 Y = = = 50 (X t − X )(Yt − Y ) = 106 aˆ = 50 − (3.533)(12) = 7.60
t =1 n 10 t =1 n 10 t =1
ORDINARY LEAST SQUARES
(OLS) Estimation Example
n
X t 120
n = 10 X = = = 12
t =1 n 10
n
Yt 500
Y =
n n

X
t =1
t = 120 Y
t =1
t = 500
t =1 n
=
10
= 50

n
106
(X
t =1
t − X ) = 30
2 ˆ
b=
30
= 3.533

(X
t =1
t − X )(Yt − Y ) = 106 aˆ = 50 − (3.533)(12) = 7.60
TESTS OF SIGNIFICANCE

Standard Error of the Slope Estimate

sbˆ =
 (Yt − Y )
ˆ 2

=
e 2
t

( n − k ) ( X t − X ) 2
(n − k ) ( X t − X )2
TESTS OF SIGNIFICANCE
Example Calculation
Time Xt Yt Yˆt et = Yt − Yˆt et2 = (Yt − Yˆt )2 ( X t − X )2
1 10 44 42.90 1.10 1.2100 4
2 9 40 39.37 0.63 0.3969 9
3 11 42 46.43 -4.43 19.6249 1
4 12 46 49.96 -3.96 15.6816 0
5 11 48 46.43 1.57 2.4649 1
6 12 52 49.96 2.04 4.1616 0
7 13 54 53.49 0.51 0.2601 1
8 13 58 53.49 4.51 20.3401 1
9 14 56 57.02 -1.02 1.0404 4
10 15 60 60.55 -0.55 0.3025 9
65.4830 30

n n n  (Y − Yˆ ) 2
65.4830
 e =  (Yt − Yˆt )2 = 65.4830 (X sbˆ = = = 0.52
t
2
− X ) = 30
2

t =1
t
t =1 t =1
t
( n − k ) ( X − X )
t
2
(10 − 2)(30)
TESTS OF SIGNIFICANCE
Example Calculation
n n

 t  t t = 65.4830
e 2

t =1
= (Y − Yˆ ) 2

t =1
n

 t
( X
t =1
− X ) 2
= 30

sbˆ =
 (Yt − Y )
ˆ 2

=
65.4830
= 0.52
( n − k ) ( X t − X ) 2
(10 − 2)(30)
TESTS OF SIGNIFICANCE
Calculation of the t Statistic

bˆ 3.53
t= = = 6.79
sbˆ 0.52

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

Critical Value at 5% level =2.306
 TESTS OF SIGNIFICANCE
Decomposition of Sum of Squares

Total Variation = Explained Variation + Unexplained Variation

 (Yt − Y ) =  (Y − Y ) +  (Yt − Yt )
2 ˆ 2 ˆ 2
TESTS OF SIGNIFICANCE
Decomposition of Sum of Squares
TESTS OF SIGNIFICANCE
Coefficient of Determination

R2 =
Explained Variation
=
 (Yˆ − Y ) 2

TotalVariation  t
(Y − Y ) 2

373.84
R =
2
= 0.85
440.00
TESTS OF SIGNIFICANCE
Coefficient of Correlation

r = R 2 withthe sign of bˆ

−1  r  1

r = 0.85 = 0.92
MULTIPLE REGRESSION ANALYSIS

Model: Y = a + b1 X 1 + b2 X 2 + + bk ' X k '

MULTIPLE REGRESSION ANALYSIS

Adjusted Coefficient of Determination

(n − 1)
R 2 = 1 − (1 − R 2 )
(n − k )
MULTIPLE REGRESSION ANALYSIS

Analysis of Variance and F Statistic

Explained Variation /(k − 1)

F=
Unexplained Variation /(n − k )

R 2 /(k − 1)
F=
(1 − R 2 ) /(n − k )
 PROBLEMS IN REGRESSION
ANALYSIS
• Multicollinearity: Two or more explanatory variables are highly correlated.
• Heteroskedasticity: Variance of error term is not independent of the Y variable.
• Autocorrelation: Consecutive error terms are correlated.
DURBIN-WATSON STATISTIC

Test for Autocorrelation

n

 t t −1
( e − e ) 2

d= t =2
n

t
e 2

t =1

If d=2, autocorrelation is absent.

TIME-SERIES ANALYSIS

• Secular Trend
– Long-Run Increase or Decrease in Data
• Cyclical Fluctuations
– Long-Run Cycles of Expansion and Contraction
• Seasonal Variation
– Regularly Occurring Fluctuations
• Irregular or Random Influences
TREND PROJECTION
• Linear Trend:
St = S0 + b t
b = Growth per time period
• Constant Growth Rate
St = S0 (1 + g)t
g = Growth rate
• Estimation of Growth Rate
lnSt = lnS0 + t ln(1 + g)
SEASONAL VARIATION

Ratio to Trend Method

Actual
Ratio =
Trend Forecast
Seasonal Average of Ratios for
=
Adjustment Each Seasonal Period
Adjusted Trend Seasonal
Forecast = Forecast Adjustment
SEASONAL VARIATION
Ratio to Trend Method:
Example Calculation for Quarter 1
Trend Forecast for 1996.1 = 11.90 + (0.394)(17) = 18.60
Seasonally Adjusted Forecast for 1996.1 = (18.60)(0.8869) = 16.50

Trend
Year Forecast Actual Ratio
1992.1 12.29 11.00 0.8950
1993.1 13.87 12.00 0.8652
1994.1 15.45 14.00 0.9061
1995.1 17.02 15.00 0.8813
Seasonal Adjustment = 0.8869
MOVING AVERAGE FORECASTS

Forecast is the average of data from w

periods prior to the forecast data point.

w
At −i
Ft = 
i =1 w
THE EFFECT OF THE PARAMETER N

• A smaller w makes the forecast more

responsive
• A larger w makes the forecast more

stable
EXPONENTIAL SMOOTHING
FORECASTS
Forecast is the weighted average of of
the forecast and the actual value from
the prior period.

Ft +1 = wAt + (1 − w) Ft

0  w 1
THE EFFECT OF THE PARAMETER 

• A smaller w makes the forecast more

stable
• A larger w makes the forecast more

responsive
AUTOGREGRESSIVE MODELING
• Used for Forecasting
• Takes Advantage of Autocorrelation
– 1st order - correlation between consecutive values
– 2nd order - correlation between values 2 periods apart
• Autoregressive Model for pth order: Random
Error

Yi = b0 + b1Yi −1 + b2Yi −2 + • • • + bpYi − p + ei

 AUTOREGRESSIVE MODELING
STEPS
• 1. Choose p:
• 2. Form a series of “lag predictor” variables
Yi-1 , Yi-2 , … Yi-p
• 3. Use Excel to run regression model using
all p variables
• 4. Test significance of Bp
– If null hypothesis rejected, this model is selected
– If null hypothesis not rejected, decrease p by 1 and repeat
your calculations
ROOT MEAN SQUARE ERROR

Measures the Accuracy

of a Forecasting Method

RMSE =
(A − F )
t t
2