Page 5 - Azerbaijan State University of Economics
P. 5
Asif Alıyev: Forecasting of Engel Curve Components with the Application of ARIMA Method
Kumar & et.al (2008) showed that food consumption deprivation index has very little
correlation with the traditional measures of poverty. They raised a question how to
combine Engel curves for some other vital goods such as health services, education,
water, energy, etc. into a single index of poverty. Almas (2012) estimated households
Engel curves for nine different countries, and observed stability of the Engel law
across the countries. James & et.al (2012) estimated food elasticities at different levels
of expenditure indicating that as income and overall food expenditure increases,
expenditure for foods will increase at a declining rate in line with Engel’s law.
Çağlayan (2012) estimated Engel curves for food and clothing in Turkey using
different econometric models indicating that the food expenditure is the largest
expenditure and the portion of this expenditure in the household budget decreases as
the income increases that is consistent with the Engel's law. Pritchett & Spivack
(2013) introduced a new simple, intuitive appeal and consistent with the Engel’s Law
ratio measure for comparison of consumption possibilities over countries using
average food shares and taking into account purchasing power of currencies and
adjusted currency conversions. Gibson & Kim (2015) revealed that food shares vary
with relative prices, but sometimes spatial price survey is not possible, and unit values
are sometimes used as price proxy. Önder (2017) studied that the consumer
consumption patterns can vary based on the geographical and socio-economic
structure differences, the survey time and the specific country circumstances. Li
(2021) empirically observed that richer consumers purchasing a larger variety of
products than poor ones. Divergent Engel curve slopes rely on the relative price and
transaction cost marginally and affect the distribution of variety gains and the
measurement of factual welfare. Laborda et. al (2021) proposed the computation of
total household expenditure in Household Budget Surveys (Engel curves) using a
generalized linear model estimator, to deal with the heteroskedasticity problem
encountered in the ordinary least squares method.
In our approach the following factors that affect Engel curve estimations were not
taken into account:
- national characteristics (mentality, psychology, religious attitudes etc.),
- personal preferences and consumer behavior,
- different income groups behaviors,
- inflation, purchasing power and currency conversions,
- economical, geographical and regional differences.
Research objective of the paper is to forecast Engel curve components with the
application of ARIMA time series forecasting method. The Engel curve evaluation is
recurrently important to get information about average wellbeing situation in the
country.
5