Econometric model of iron ore through principal component analysis and multiple linear regression

Abstract

Price of iron ore is affected by instabilities of microeconomic balance between supply and demand. Periods of equilibrium adjustment result in huge swings, growth or global recession. They also impact the viability of mineral enterprises and generate consequences to important global economic scenarios. This research aims to evaluate the market variables capable of influencing the price of iron ore through multivariate statistical techniques. Principal component analysis and multiple linear regression, booth multivariate statistical techniques were used. The studied variables were rate export of iron ore and concentrates from Brazil, steel production from China, steel production from Japan, production from Europe, steel production from the United States, steel production from India, steel price, coal price, China’s Construction Gross Domestic Production, United States construction index, oil price and global oil production. First three components explained 89.12% of the variability of the data matrix. Multiple linear regression highlighted the significance of five variables. They are export iron ore from Brazil, steel production from China, price of coal, steel production from India and price of steel.

Key words
multivariate statistics; iron ore; principal component analysis; multiple linear regression; mineral economy