Research Studies Project Report (M.B.A.) - Asian Institute of Technology, 1998

With English abstract

Memoire de recherche soum1s pour l'obtention du diplOme de Master of Business Administration, School of Management

The forecasts are projections that are based on the flow pattern of past figures. These are highly important tools for management and line personnel in their daily course of activities and decision making. The purpose of this research is to show managers in Lao companies some of the methods which may help them to improve their forecasts. This research follows three main principles: first, that the level of mathematics and statistics has been kept to that considered as appropriate to people whose interest is in obtaining forecasts for practical purpose. Secondly, that the emphasis is on the simpler and more widely used procedures, with some modifications provided that the level of precision must be satisfactory, using sales data of LAO BREWERY CO., LTD as the case study. The third principle is the critical approach that is essential in selecting the most suitable method from the many which are available. The test of data shows that the past sales figures of LAO BREWERY CO., LTD behave randomly with an upward trend, and have no seasonality. The direct use of Moving Average, Exponential Smoothing, and Linear Regression is inefficient because they cannot keep track of the presence of randomness. The corrected methods suggested in this research seem to give good results. Holt-Winters, Decomposition, and Box-Jenkins methods are not recommended due to their complexities and seem not accessible to Lao managers' skills. The research has dealt with short-term forecasts and time senes, while the long-term forecasts and causal methods have not been taken into account.