Business Cycle Forecasting Of Pharmaceutical Retail Sales in the United States

Abstract

The competitive climate in the B2B segment of the pharmaceutical industry is ever increasing. The number of patents reaching their expiration date will climb to new heights during the next three years. Along with this, government regulations have contributed to decreasing profit margins over the past 8- 10 year. Furthermore, the pharmaceutical industry has experienced a fundamental change in its nature and is suddenly seeing drops in growth rates, which were only last experienced since before the Second World War. The need for a forecasting model that can predict fluctuations in the pharmaceutical sales has therefore never been greater. A review of existing theoretical and empirical contributions to the understanding of business cycles has been conducted, and key findings from these reviews have been implemented in a specially constructed model. Using this model as a foundation, several indicators from the macro-economy, various industries and the pharmaceutical industry itself have been analyzed with the purpose of exposing a forecasting potential. The thesis bases its selection of indicators on a thorough chronological review of both early and modern business cycle theory, as well as a profound understanding of methodological views that have helped define the criteria for the collection of data. This results in a broad spectrum of indicators, which is optimal according to several theorists that highlight diversity among the indicators as essential for forecasting. In order to get the most nuanced and complete view of the total business cycle fluctuations in the economy, more than 40 macroeconomic indicators were tested. A recurring argument is that no two industries are alike and that the selection, processing and interpretation of relevant indicators cannot be done in a general fashion. In accordance with the chosen methodology and on a trial and error basis, the selection of indicators was narrowed down and 22 indicators were selected to be included in the model. The model was constructed specifically to forecast fluctuations in the pharmaceutical retail sales in the United States. Based on the calculations of the MACD method, the indicators were all compared to a main indicator, which resulted in an insight into each indicator’s forecasting potential. It was quickly found that the standard of measure for forecasting, the Conference Board Leading Index, was not suited for the special nature of the pharmaceutical industry, which created a demand for a revised composite index. Such an index was constructed based on the analytical results of each indicator, which included lead/lag times, Cross Correlation Coefficient, standard deviation and number of false signals. There are carried out a robustness analysis which confirms the correlations found in the American economy, this can imply a general relationship which could be used by pharmaceutical companies in America.