Analysis of the properties of weather regressors for econometric modelling: Example of weather stations in Poland

Abstract

Many recent works indicate the existence of a significant relationship between weather factors such as pressure, humidity, windspeed, sum of falls or sunshine and rates of return for stocks quoted on stock exchange. A properly conducted econometric study requires a careful analysis of the properties of the factors that will be used in the econometric model to explain the development of the dependent variable. The aim of the research is to check if the weather factors could be used as econometric regressors by verifying their statistical propensities. The analysis was held on the weather stations located in eight cities in Poland: Poznan, Kolo, Plock, Warsaw, Wroclaw, Opole, Katowice and Rzeszow. These cities host registered offices of the biggest companies of the energy sector in Poland. The research methods were focused around basic statistics and normality tests of distributions of weather factors' (four types), as well as the autocorrelation of regressors.