Study of Economic Systems Using the Simulation-Based Statistical Modeling Method

Abstract

This article analyzes theoretical and methodical aspects of study of economic systems using the simulation-based statistical modeling method. Principles of simulation-based statistical modeling are demonstrated on “coin toss” and “improved Buffon experiment” simple classic models realized within the environment of MS Excel tabular processor. The theoretical basis of the method – the law of large numbers and the central limit theorem – are also demonstrated by the way of computer experiments. Setting of adequate laws of random distribution is the main condition for efficient application of simulation-based statistical modeling method. Wrong assumption regarding the type of distribution laws may result in significant errors in assessment of the output parameters of the system under consideration. Selection of a distribution law type should be based on the economic essence of the modeled random variable. The logical scheme and phases of the simulation-based statistical modeling are demonstrated by the example of investment project’s risk assessment. The project cash flow model is based on known estimated NPV ratios. External and internal random factor impact model allows to set various laws of random distribution, and to take into account correlations between them. Statistical processing model, together with the module for obtaining of typical statistical estimates, includes the module for construction of empirical distribution function and selection of projects with NPV close to the modal value within the set interval. The modeling process shows that the obtained mean NPV is too optimistic in certain situations, and it would be reasonable to take the modal value of NPV as its projected value. The structure of the computer model developed using the MS Excel tabular processor environment is considered.

DOI: 10.5901/mjss.2015.v6n4s4p369