Predicting Steel Consumption for India using Multi-Variate Regression Analysis of Data from Similar Countries

Abstract

An attempt is made to create a statistical model for predicting finished steel consumption for India in the medium term by studying historical data from similar sized steel economies. Economies were selected based on their size and steel consumption profile. These were further narrowed down to those economies where reliable economic data was available for a per capita GDP range where India has been in the recent past and would be in the medium term. Finally the data for China and South Korea were found suitable. The data was split in periods where said economies were in the per capita GDP range of $500-2000 and $2000-10,000 (constant 2021 US$). India is currently at a per capita GDP of $2000. Thereafter, utilising regression analysis, starting with a single variable regression and advancing into regularised multi-variate regression, an attempt was made to narrow down the economic predictors for steel consumption. Finally a model was derived after multiple rounds of data fitting that could predict the per capita steel consumption for India. To arrive at a steel consumption number for a particular year, the economic variables responsible for steel consumption were forecasted for the year and a range was arrived at. The resultant range for steel consumption for India in 2030 is 157-188 mn MT.