Rounding Errors and their Impact on the Accuracy and Reliability of Statistical, Financial and Economic Calculations

Abstract

Statistical observations and calculations are inevitably accompanied by errors of both systematic and random nature. Rounding errors occur both in the process of observation and in obtaining and interpreting the final results. Methods for estimating, minimizing and eliminating rounding errors can be developed based on a number of well-known provisions from the theory of measurement errors. Rounding errors do not follow the law of normal distribution. The article provides an assessment of the maximum and minimum permissible absolute and relative rounding errors that ensure the required or specified accuracy and reliability of the final results when carrying out statistical and financial-economic calculations at enterprises, organizations, and during training sessions in universities in the course “Economic Statistics”. The article substantiates and gives recommendations on setting the limit values of numbers rounded off with allowable or specified absolute and relative errors. The corresponding tables have been compiled and proposed, and an example of calculations in compliance with the developed recommendations is given. The results of the work can be useful in carrying out balance sheet calculations and preparing financial statements.

Keywords

• properties of random errors
• absolute and relative rounding errors
• maximum rounding errors
• relation between maximum and mean squared error
• maximum allowable rounding errors
• minimum limits of rounded numbers allowable values