CASESTUDY

Estimating Demand for Dabur HONEY by Dabur Nepal Pvt. Ltd.

In 2017, Dabur Nepal Pvt. Ltd., initiated an empirical estimation of demand for its HONEY
product. The Dabur Nepal is formulating pricing and promotional plans for the coming year, and
management is interested in learning how pricing and promotional decisions might affect sales.

Dabur Nepal Pvt. Ltd., has been marketing HONEY for several years, and its market research
department has collected quarterly data over two years for six important marketing areas,
including sales quantity, the retail price charged for the HONEY, local advertising and
promotional expenditures, and the price charged by a major competing brand of
PATAHJAI_HONEY. Statistical data published by a research paper on population and
disposable income in each of the six market areas were also available for analysis. It was
therefore possible to include a wide range of hypothesized demand determinants in an empirical
estimation of HONEY demand. These data appear in Table 1.

The following regression equation was fitted to these data:

Qit = α + β 1Pit + β 2Ait + β 3PCit + β 4 Yit + β 5POPit + β 6TVit + uit

P_PR A_ADVE PC_Patanja POP_Pop’ TV_Tim

Q_U Y_Income
Year ICE XP li_Competit n e
NIT Thousand
City Rs. Thousand or Price Rs. Thousand Variable
Biratnager 2017_4 27500 550 10 375 41.5 2650 8
A 2017_3 25000 600 7.5 375 40.5 2500 7
  2017_2 25000 575 10 375 40 2450 6
  2017_1 25000 575 5 400 39.5 2350 5
  2016_4 27500 525 10 400 39.5 2300 4
  2016_3 22500 500 7.5 325 39 2250 3
  2016_2 25000 525 7.5 375 39.5 2150 2
  2016_1 22500 600 5 425 39.5 2150 1
Janakpur 2017_4 27500 600 5 425 40 2300 8
B 2017_3 25000 600 5 400 39.5 2300 7
  2017_2 27500 525 10 425 39.5 2250 6
  2017_1 25000 550 5 400 39 2200 5
  2016_4 22500 600 5 400 39 2250 4
  2016_3 22500 550 5 375 39 2200 3
  2016_2 22500 625 5 400 39 2200 2
  2016_1 22500 600 7.5 400 38.5 2150 1
Chitwan 2017_4 32500 600 5 400 46 6200 8
C 2017_3 32500 550 15 375 45.5 6150 7
  2017_2 27500 600 5 375 45 6100 6
  2017_1 22500 600 10 350 44.5 6150 5
  2016_4 30000 550 5 375 45 6200 4
  2016_3 30000 575 15 350 44.5 6250 3
  2016_2 25000 600 5 450 44.5 6100 2
  2016_1 27500 575 5 375 44 6050 1
Bhirawa 2017_4 35000 500 15 400 47.5 1600 8
D 2017_3 32500 575 10 400 47 1650 7

1|Page
   2017_2 32500 550 7.5 425 47 1600 6
  2017_1 30000 600 12.5 400 46.5 1550 5
  2016_4 27500 550 5 350 46 1550 4
  2016_3 25000 600 5 325 46.5 1500 3
  2016_2 27500 575 10 350 47 1450 2
  2016_1 30000 550 10 425 46.5 1450 1
Nepaljaug 2017_4 17500 600 2.5 375 35.5 300 8
E 2017_3 17500 625 2.5 375 35 290 7
  2017_2 15000 600 5 375 34.5 285 6
  2017_1 17500 575 2.5 375 34.5 270 5
  2016_4 15000 625 2.5 325 34 265 4
  2016_3 17500 575 2.5 375 34 270 3
  2016_2 15000 575 5 350 34 275 2
  2016_1 17500 575 2.5 400 34 280 1
Dhangadi 2017_4 27500 625 5 400 46 1500 8
F 2017_3 27500 625 12.5 350 46 1450 7
  2017_2 27500 625 5 450 45 1300 6
  2017_1 25000 625 5 375 44.5 1450 5
  2016_4 30000 550 7.5 425 44.5 1350 4
  2016_3 30000 575 12.5 425 44 1100 3
  2016_2 27500 600 12.5 400 43.5 1050 2
  2016_1 25000 575 10 400 43.5 1025 1

Dependent Variable: UNIT (Q)

Variable Coefficient Std. Error t-Statistic Prob.  
C -5084.553 5048.082 -1.007225 0.3197
PRICE (P) -35.61085 7.031919 -5.064173 0.0000
ADVEXP (A) 208.7423 77.49861 2.693498 0.0102
COMPPR (PC) 38.25608 7.153296 5.348035 0.0000
INCOME (Y) 781.5315 66.39043 11.77175 0.0000
POPN (POP) 0.261445 0.125639 2.080916 0.0437
TV (TV) 351.7874 92.46384 3.804595 0.0005
R-squared 0.926903     Mean dependent var 25312.50
Adjusted R-squared 0.916206     S.D. dependent var 4990.017
S.E. of regression 1444.474     Akaike info criterion 17.52292
Sum squared residual 85546693     Schwarz criterion 17.79580
Log likelihood -413.5500     Hannan-Quinn criterion 17.62604
F-statistic 86.64936     Durbin-Watson stat 2.199748
Prob(F-statistic) 0.000000

AIC= Optimum lag selection. Lowest AIC value better the model. That is the guide line.
Schwarz criterion

Q is the quantity of pies sold during the tth quarter, P is the retail price in rupees of Dabur
HONEY; A represents the rupees (in thousands) spent for advertising and promotional activities;
PC is the price, measured in rupees, charged for competing PATANJALI_HONEY; Y is
thousands of Rs. of disposable income per household; POP is the population of the market area
(in thousands of persons); and T is the trend factor (2016-1 = 1, 2017-4 = 8). The subscript i

2|Page
indicate the regional market from which the observation was taken, whereas the subscript i
represents the quarter during which the observation occurred. Least squares estimation of the
regression equation on the basis of 48 data observations, or eight quarters of data for each of six
areas, resulted in the estimated regression coefficients and other statistics given in Table 1.

The individual coefficients for the Dabur HONEY demand regression equation can be
interpreted as follows. The intercept term, -5084.553, has no economic meaning in this instance;
it lies far outside the range of observed data and obviously cannot be interpreted as the demand
for Dabur HONEY, when all the independent variables take on zero values. The coefficient for
each independent variable indicates the marginal relation between that variable and sales of
HONEY, holding constant the effect of all the other variables in the demand function. For
example, the -35.61coefficient for P, the price charged for Dabur HONEY, indicates that when
the effects of all other demand variables are held constant, each Rs.10.00 increase in price causes
quarterly sales to decline by roughly 36 liters. Similarly, the 208.74 coefficient for A, the
advertising and promotion variable, indicates that for each Rs.1,000 (one-unit) increase in
advertising during the quarter, on an average 209 liters additional HONEY are sold. The 38.256
coefficient for the competitor-price variable indicates that demand for Dabur HONEY rises by
on an average 38 liters with every Rs.10.00 increase in competitor prices. The 781.53 coefficient
for the Y variable indicates that, on average, a Rs.1,000 (one-unit) increase in the average
disposable income per household for a given market leads to 782 liters increase in quarterly
HONEY demand. Similarly, a 1,000 person (one-unit) increase in the population of a given
market area leads to a small 0.261 liters increase in quarterly HONEY demand. Finally, the
351.787 coefficient for the trend variable indicates that HONEY demand is growing in a typical
market by on an average 352 liters per quarter. This means that Dabur Nepal is enjoying secular
growth in HONEY demand, perhaps as a result of the growing popularity of Dabur products of
HONEY in general.
Individual coefficients provide useful estimates of the expected marginal influence on demand
following a one-unit change in each respective variable; however, they are only estimating. For
example, it would be very unusual for a Rs.10.00 increase in price to cause exactly a -33.985-
unit change in the quantity demanded. The actual effect could be more or less. For decision-
making purposes, it would be helpful to know if the marginal influences suggested by the
regression model are stable or instead tend to vary widely over the sample analyzed.

In general, if it is known with certainty that Y = a + bX, then a one-unit change in X will always
lead to a b-unit change in Y. If b > 0, X and Y will be directly related. If b < 0, X and Y will be
inversely related, if no relation at all holds between X and Y, then b = 0. Although the true
parameter b is unobservable, its value it estimated by the regression coefficientb^ . If b^ = 10, a 1-
unit change in X will increase Y by 10 units. This effect may appear to be large, but it will be
statistically significant only if it is stable over the entire sample. To be statistically reliable, b^
must be large relative to its degree of variation over the sample.

In a regression equation, there is a 68-percent probability that b lies in the interval b^ ± one
standard error (or standard deviation) of the coefficientb^ . There is a 95- percent probability that
b is lies in the interval b^ ± two standard errors of the coefficient. There is a 99-percent

3|Page
probability that b is in the interval b^ ± three standard errors of the coefficient. When a coefficient
is at least twice as large as its standard error, one can reject at the 95-percent confidence level the
hypothesis that the true parameter b equals zero. This leaves only a 5-percent chance of
concluding incorrectly that b ≠ 0 when in fact b= 0. When a coefficient is at least three times as
large as its standard error (standard deviation), the confidence level rises to 99 percent and
chance of error falls to 1 percent.

A significant relation between X and Y is typically Indicated whenever a coefficient is at least

twice as large as its standard error; significance is even more likely when a coefficient is at least
three times as large as its standard error. The independent effect of each independent variable on
sales is measured using a two-tail statistic where:
b^
t−statistics=
Standered error of b^

This t-statistic is a measure of the number of standard errors between b^ and a hypothesized value
of zero. If the sample used to estimate the regression parameters is large (for example, n > 30),
the t-statistic follows a normal distribution, and properties of a normal distribution can be used to
make confidence statements concerning the statistical significance of b^ . Hence t = 1 implies 68
percent confidence, t = 2 implies 95-percent confidence, t = 3 implies 99-percent confidence, and
so on. For small sample sizes (for example, d.f. = n - k < 30), the t distribution deviates from a
normal distribution, and a t table should be used for testing the significance of estimated
regression parameters.

Another regression statistic, the standard error of the estimate (S.E.E.), is used to predict values
for the dependent variable given values for the various independent variables. Thus, it is helpful
in determining a range within which one can predict values tor the dependent variable with
varying degrees of statistical, confidence. Although the best estimate of the value for the
dependent variable is Y^ , the value predicted by the regression equation, the standard error of the
estimate can be used to determine just how accurate this prediction Y ^ is likely to be.

Assuming that the standard errors are normally distributed about the regression equation, there is
a 68-percent probability that actual observations of the dependent variable Y^ will be within the
range Y^ ± one standard error of the estimate. The probability that an actual observation of Y will
lie within two standard errors of its predicted value increases to 95 percent. There is a 99-percent
chance that an actual observed value for Y will lie in the range Y ^ ± three standard errors.
Obviously, greater predictive accuracy is associated with smaller standard errors of the estimate.

Mrs. Smyth’s could forecast the total demand for its pies by forecasting sales in each of the six
market areas, then summing these area forecasts to obtain an estimate of total pie demand. Using
the results from the demand estimation model in Table 1 and data from each individual market
from Table 2, (Result of estimation) it would also be possible to construct a confidence interval
for total pie demand based on the standard error of the estimate.

a) Describe the statistical significance of each individual independent variable included in

the Mrs. Smyth’s frozen fruit pie demand equation.

4|Page
b) Interpret the coefficient of determination ( R 2 ) for the Mrs. Smyth’s frozen fruit pie
demand equation.
c) What is the expected value of next quarter’s unit sales in the Baltimore, Md., market?
d) To illustrate use of the standard error of the estimate statistic, derive the 95-percent
confidence interval for next quarter’s actual unit sales in the Baltimore, Md., market.

5|Page