An Econometric Analysis of Demand Pattern for
Major and Minor Durables
Dr. Prasant Sarangi
School of Management and Computer Science Apeejay Institute of Technology
Abstract: Demand for durables depends not only on the ability of consumers to purchase but also ebbs and flows of confidence of consumers. The households acquire some durables to maintain physical comforts or needs; some associated with tradition, culture and habit; some for pleasure and some others for maintaining social status. So the purchase behaviour of durables is quite different from that of non-durables. The effect of some important quantitative determinants on the demand for major and minor durables has been estimated from Logit model by maximum likelihood method using household level consumption expenditure data. It is observed that income plays a vital role in the purchase of both major and minor durable. In addition to this, the number of durable possessed and education level of the household head have significant positive effect on the demand for these durables.
Key Words: Durable goods, Engel function, Logit, Household, Ownership
Demand for durables has certain peculiar characteristics. When a new durable is introduced initially a small fraction of the population will purchase it and gradually its potential buyers will increase. This will continue till all the potential buyers own it. Ultimately it becomes an established commodity or otherwise called as branded product. This is what we call the diffusion of durables, which may be due to the outcome of the interaction of a number of quantitative and qualitative factors. Some durables become outdated which are no longer in demand due to availability of better substitutes. The very nature of infrequent purchase behaviour of durables paves the way for separate studies on pattern of demand for durables. Empirical studies on durables are poorly represented whereas non-durables have been intensively studied . The recent literature shows that such studies are scanty and a few studies done so far have estimated the Logit or Probit model to compute threshold probability, range of predicted probabilities and threshold level of income for the acquisition of durables.
II. REVIEW OF RELATED LITERATURES
incorporating some important determinants of demand for major and minor durables. He has explored the effect of some of the important determinants of demand for durables; estimated the threshold level of income and probabilities for acquiring durables and diffusion of these durables among different occupation group of households in Orissa. Sarangi et al.  has estimated the diffusion of major and minor durables by using both qualitative and quantitative determinants in a business town of Odisha state. The result proved various factors affecting on the demand of durables. In other studies Sarangi et. al.,  and Srivastave and Sarangi  have used econometrics tools in modeling daily returns in capital market.
There is a growing body of literature on time-series and cross-section studies on durables in marketing research but we rarely find studies on det of demand for household durables. To bridge this gap, an attempt has been made in this study to explore the effect of some important quantitative and qualitative determinants of demand for durables and the threshold level of income and probabilities for acquiring each of the durables and diffusion of these durables among different occupation group.
III. OBJECTIVE OF THE STUDY
The basic objective of the present study is to explore the effect of various determinants in the purchase of major and minor durables.
i) To explore the effect of some important quantitative determinants of demand for major durables in the study area
ii) To explore the effect of some important quantitative determinants of demand for minor durables in the study area
IV. METHODOLOGY OF THE STUDY
Since household level data on different aspects of demand for durable goods is not available, the study is based on primary household level micro data collected from household survey.
4.1 The Data
It is an empirical study based on primary data collected during the first quarter of 2014 (January to March. The primary data was collected by a sample survey of 300 households of Greater Noida, a small town of Gautam Budh District of Uttar Pradesh state. The area of investigation has been divided into three segments on the basis of occupation of the head of the household. Segment-1 basically constitutes households of government and semi-government officials, Segment-2 is a concentration of business community and finally, Segment-3 constitutes working class households such as carpenters, masons and other manual workers. The list of households of the three segments was prepared separately for each Segment and 100 households were selected at random (using random table) from each segment of the town. For the collection of data, household expenditure survey schedules specially prepared for the purpose were canvassed by direct personal interview with the head of the household.
between Rs.500/- to Rs. 5000/-, it is considered as minor durable and more than Rs. 5000/- then it is considered as major durable. We have considered here only those major and minor durables possessed by most of the sample households. Major durables include television, refrigerator, two wheeler and air cooler and minor durables include almirah, gas stove, pressure cooker; bicycle, electric fan, radio and costly watch.
4.2 Quantitative determinants of demand for durables
Income is one of the important determinants of demand for durables. Average monthly household expenditure is used as a proxy variable for average monthly household income. The average monthly disposable income of a household has a strong positive effect on the acquisition of durables and household expenditure on durables. Household size and number of durables possessed by the household has negative effect on the demand of durables and household expenditure on durables. Age, sex and level of education of the household head are other important determinants in the demand of durables. Younger household heads are generally more active in the acquisition of durables than the older ones. In this section, we have estimated the logit model separately for major and minor durables to know the effect of some of the important quantitative determinants of demand for durables. The stochastic specifications of logit models for major and minor durables are:
Yij* are the qualitative choice dependent variables forithmajor and minor durable respectively of
thejthhousehold. Than Y=1, if the household owns a major/ minor durable and Y=0 if it does not. The determinants
of demand for ith minor and major durables are the income of the household,
Xij; household size,
Nij; number of
durables possessed by the household,
Dij; age of the household head,
Aij; years of schooling,
Eij is used as a
proxy variable for the level of education of the household head;
are the parameters; and
the stochastic terms of these models. The parameters of the Logit models are estimated by maximum likelihood method. Ownership probabilities and threshold level income for acquiring each of the major and minor durables are also estimated from these models. The extent of diffusion of the major and minor durables is estimated from these models using the procedure of Ironmonger (1972).
V. EMPIRICAL ANALYSIS
household. The parameter estimates of the index function of logit model estimated for four major durables is summarized in Table-1 and for seven minor durables is summarized in Table-2.
The estimates of the index function of logit model for major durables represented in Table-1 reveals that income, number of durables possessed, age and education level of the household head have significant effect on the demand for most of the major durables. Income as one of the most important determinants has significant positive effect on the demand for all the major durables as expected. Household size which is another important variable has significant negative effect on the demand for television and two-wheeler but insignificant positive effect on the demand for refrigerator and air cooler. The reason for negative household size effect may be addition of a child which enhances household expenditure thereby reducing expenditure in the acquisition of less useful major durables. Households are interested for the better education of their children and spend more on education and less on durables. Age of the head of the household has significant negative effect on the demand for two wheelers and also on refrigerator but significant positive effect on the demand for television and air cooler which may be due to the fact that older household heads do not prefer to own two wheelers and do not prefer to eat food and vegetable preserved in refrigerators. So, household size and age of the household head have no definite effect on the demand for major durables. It is also noticed that the number of durables possessed and education level of the household head have significant positive effect on the demand for major durables which may be due demonstration effect and availability of better substitutes. Most of the coefficients of the logit model for major durables are found to be significant at 1 per cent level of significance.
The likelihood ratio (LR) and Chi-square test which are generally used to test the joint significance of slope of coefficients of the Logit index function, are also found to be significant for all the four major durable goods.
R2), a measure of goodness of fit is found to be the highest for refrigerator and the lowest for
air cooler. Even though the value of
RMcF2 is not found to be so high for most of the major durables but
also considered as an acceptable indicator to measure the goodness of fit is found to be more than 0.60 for most of the major durable.
steel almirah and negative effect in the purchase of radio, gas stove, watch and bicycle. On the other hand, education level of the household head has positive significant effect on most of the minor durables except watch and bicycle. The ownership of gas stove and watch are in no way helpful in the decision making process of the household. It is
observed that the Likelihood Ratio
and values of2(chi-square) are significant for most of the minor durables
except radio and watch. Most of the coefficients of radio and watch are found to be not significant as the two minor
durables are fully diffused and most of the sample households possess these two minor durables. The
RMcF2 is also
found to be high for all minor durables except watch and radio but the percentage of prediction as reflected by count
Ris very high for watch (95 per cent) and it is very low for radio.
VI. SUMMARY OF THE FINDINGS
It is concluded that the determinants like income, the number of durables possessed, level of education of household head have significant positive effect on the demand for most of the major and minor durables. Income is one of the important determinants which has significant positive effect on the demand for the major and most of the minor durables except watch and pressure cooker. Household size and age of the household head have no define pattern of demand for durables in the sense that they have positive effect on the demand for some major and minor durables and negative effect for others. But level of education of the household head has positive significant effect on the demand for major durables and most of the minor durables except watch and bicycle. Similarly, the number of durables possessed by a household also has positive effect on the demand for durable goods.
 Harberger, A. “The Demand for Durable Goods”. University of Chicago Press: Chicago, 1960.
 Dernburg, T.F. “Consumer Response to Innovation: Television”.Cowles foundation. Discussion paper No. 121, 1957.
 Cramer, J. S. “A Statistical Model of the Ownership of Major Consumer Durables”.Cambridge University Press: Cambridge, England, 1962.  Patanaik, S. “Demand for Major Household Durables in Orissa”,M.Phil. Dissertation (Unpublished) submitted to Berhampur University,
Berhampur, Orissa, 2004.
 Sarangi, P., S. Pattanaik, and B.K. Panda. “Demand for Major and Minor Durables In Orissa: An Econometrics Analysis”.Vision, Vol. 17, No. III, 2008
 Sarangi, P., Shashank Dublish, and Ankur Srivastava. “Forecasting Performances of GARCH Families of Models”.Apeejay-Journal of Management Sciences and Technology, Vol.-1, Issue-1, October, 2013.
 Srivastava, A. and Prasant Sarangi. “Are GARCH Specifications Superior among GARCH Types of Models in Estimating Financial Volatility? An Experiment”.Apeejay-Journal of Management Sciences and Technology, Vol.-1, Issue-2, February, 2014.
 Duesenberry, J.S. “Income, Saving and Theory of Consumer Behaviour”.Haward University Press: Cambridge, Massachusetts, 1949.
TABLE – 1
PARAMETER ESTIMATES OF INDEX FUNCTION OF GENERAL LOGIT MODEL ON DEMAND FOR MAJOR HOUSEHOLD DURABLES (MODEL-1)
Income HH Size
Age Educati on
1 T.V -7.6430
(-1.989) 0.3691 * (3.168) -0.594 ** (-2.597) 0.374 * (2.706) 0.263 * (2.828) 0.321 *
(2.923) -17.518 -26.738 0.708 23.421
* 0.531 0.467 88.6
2 Two-Wheeler -3.4275 (-1.121) 0.265* (3.117) 0.271* (2.680) 0.282* (2.828) -0.675* (-3.156) 0.214* (3.167)
-26.650 -33.821 0. 891 15.412* 0.634 0.369 87.8
3 Refrigerator -2.1897 (-0.245) 0.4206* (4.295) -0.511 (-0.657) 0.384** (2.243) -0.207* (-3.937) 0.461* (3.212)
-18.315 -25.214 0.823 26.321* 0.651 0.521 93.5
4 Air Cooler -6.9324 (-1.605) 0.591* (3.604) 0.344 (0.066) 0.971* (5.670) 0.842* (4.107) 0.286* (4.166)
-14.214 -17.102 0.831 7.792 0.296 0.231 88.7
NOTE:1. Estimated from Survey data using Limdep Econometrics Software-Version 7.03. The figures in the parenthesis are the ‘t’ ratios; LR-likelihood ratio;
)-Restricted Likelihood Function;
)-Unrestricted Likelihood Function; R2is the Coefficient of Determination, 2
R– McFadden Squared
TABLE – 2
PARAMETER ESTIMATES OF INDEX FUNCTION OF GENERAL LOGIT MODEL ON DEMAND FOR MINOR HOUSEHOLD DURABLES (MODEL-2)
(%) Income HH Size Total
Durables Age Education
1 Radio -3.402*
(-0.558) 0.254* (3.163) 0.852 (0.377) 0.135 (0.182) -0.354* (-2.701) 0.759* (2.142)
-32.321 -35.542 0.754 5.598 0.121 0.152 64.78
2 Fan -42.398
(1.962) (1.354)1.286 0.467
(2.703) -12.844 -17.569 0.856 21.601
* 0.875 0.432 97.18
3 Gas Stove -0.816 (-0.395) 0.877* (2.515) -0.489 (-0.705) 0.323 (2.197) -0.215 (-0.425) 0.728 (0.749)
-26.241 -34.423 0.881 16.421* 0.412 0.322 74.14
(Costly) (0.001)15.457 (-1.068)-0.796 (0.012)10.887 (0.145)11.678 (-0.287)-2.410 (-1.145)-2.942 -0.986 -6.712 0.325 4.586 0.113 0.279 96.55
5 Pressure Cooker -8.605 (-2.851) -0.217 (-0.869) -.6540** (-1.980) 0.715* (2.892) 0.245* (2.393) 0.264* (2.535)
-18.716 -34.231 0.657 33.521* 0.548 0.547 89.21
6 Steel Almirah -4.548 (-1.570) 0.386* (2.601) -0.254 (-0.637) 0.233* (1976) 0.448* (2.876) 0.248 (0.303)
-25.221 -30.342 0.756 13.536** 0.265 0.325 71.81
7 Bicycle 4.456
(1.002) 0.6301* (2.318) 0.614** (1.970) 0.872** (1.995) -0.321 (-1.919) -0.458* (-2.789)
-13.616 -18.889 0.865 16.718* 0.548 0.325 95.21
NOTE:1.Estimated from Survey data using Limdep Econometrics Software-Version 7.03. The figures in the parenthesis are the ‘t’ ratios; LR-likelihood ratio;
)-Restricted Likelihood Function;
Likelihood Function; R2is the Coefficient of Determination, 2