|
Abstract:
|
This presentation will explore the effect of weather on economic data, in the context of official statistics that may be analysed on a seasonally adjusted basis. We live in an era where weather extremes are becoming more pronounced and protracted than in the past. It is conceivable that long stretches of unusually low or high temperatures, prolonged drought conditions, extended rainfall, to name a few, may impact the behaviour of both individuals and businesses. Seasonal adjustment removes repeating, equally spaced patterns from the corresponding time series data, along with movement due to various holidays and other predefined regressors, but it DOES NOT remove the impact that atypical weather may have. To fill the void, many national statistical agencies have been studying the use of weather related indicators in their time series analyses, including seasonal adjustments. We discuss alternatives for incorporating the data into seasonal adjustment and ad hoc analysis. Using X12 ARIMA, we present a few examples of how weather related regressor may be incorporated into seasonal adjustment to help explain unusual movements in economic time series.
|
