Исследовательский Центр ИПМ
Kasrtyčnicki Ekanamičny Forum

TN/01/13 Deseasonalizing Belarusian GDP

Various economic indicators have a seasonal pattern and demonstrate regular yearly fluctuations. Seasonal variation makes it difficult to analyze the raw data and does not allow identifying explicitly short-term dynamics of the indicator and its long-term trends. The paper proposes a novel approach that allows conducting seasonal adjustment when several observations in analyzed time series are contaminated and substantially distort the results of seasonal adjustment. A dramatic example of such a situation is seasonally adjusted data for Belarusian real GDP in 2013Q1, when an application of standard statistical methods resulted in an annualized growth rate of about 40%.

The proposed approach allows identifying the contaminated observations and carrying out seasonal adjustment of real GDP eliminating an influence of these observations. As a result, according to our estimates, the corrected figures for Belarusian real GDP growth in 2013Q1 were 0.8% and 3.2% for quarter-to-quarter and annualized growth rates, respectively.

The paper discusses in detail the methodology of seasonal adjustment of Belarusian real GDP and proposes policy recommendations that are of practical importance for statistical agencies, central banks, economic observers and academic researchers.