Demand forecasting for sales analysis of petroleum products

Abstract

Crude oil is one of the major imports of India. In 2015, India imported crude oil worth USD 104.7 billion which accounts for about 26.8% of total India’s imports.1 The oil consumption in India is about 4.1 million barrels/ day.2 Some of the major products of petroleum that are widely consumed in the nation include High speed diesel (HSD), Motor Spirit (MS), Superior Kerosene Oil (SKO) etc. High speed diesel is mainly used a fuel in trucks, buses, diesel generators, diesel cars etc. Motor spirit or petrol is a fuel mainly used in spark ignition engines in automobiles, motorbikes etc. Superior kerosene oil is mainly used as a fuel for domestic consumption in lighting and heating purposes. Some amount of it is also used in the manufacture of insecticides and other pest control chemicals. With India’s GDP growing at a rate of 7.6%, the country’s petrol and diesel consumption is expected to rise just over 7% and 4% respectively.3 Another major reason which has boosted the consumption of oil in India is the declining crude oil prices. With the recent shale oil boom, there has been a steep increase in the production of crude oil which has overshot the supply of crude oil globally. As a result of which the prices of crude oil has plummeted from $ 109/ barrel in June 2014 to $ 47/ barrel in August 2016.4 Consequently in India there has been a rate cut in prices of petrol as many as 11 times since August 2014. Therefore due to various macroeconomic factors playing a major part in the supply and demand for oil consumption in the country it becomes crucial to forecast the demand for oil in the near future to predict sales of these products. For our project we will be predicting the demand for 3 petroleum products namely High Speed Diesel (HSD), Motor Spirit (MS) or petrol and Superior Kerosene Oil (SKO) from a time period from 2007 to 2014. In each of this data set we will be using monthly sales data from April 2007 to March 2013 as the model period to forecast the sales for 2014 and compare it with actual sales of 2014. Sales of HSD, MS and SKO will vary with respect to inflation as it would affect the buying behavior of people. So WPI is taken as one of the independent variables to predict the sales of HSD, MS and SKO. Also as the purchasing power of people increases they would tend to consume more of the fuel. Hence GDP is taken as another independent variable as a proxy for purchasing power to study its impact on the sale of HSD, MS and SKO. Likewise as Index of Industrial production and sales of cars increase, it is expected that there would be increase in the sales of HSD and MS as well Hence to understand weather or not these above mentioned independent variables influence the sales of HSD and MS, which is our dependent variables in this case, a regression analysis was run between them. The results of the regression analysis have been discussed in the coming sections.