Comparative Performance of Two Multinomial-Based Methods for Obtaining Lower Bounds on the Total Overstatement Error in Accounting Populations

Accurately estimating a lower bound on the total overstatement error in an accounting population is important in many audit decision situations. This article reports the results of a simulation study of three lower-bound estimation procedures based on a multinomial model with a dollar-unit sampling scheme. The ability of each procedure to attain a specified confidence level is assessed with respect to changes in error rate and random sampling method. The simulation study shows that the tested procedures perform well in achieving specified lower-bound coverages on the total overstatement error in the chosen accounting populations. Coauthors are Robert Plante, Kam-Wah Tsui, and P. Kannan

Econometrics · Economics · Multinomial distribution · Statistics · Forecasting Techniques and Applications · Insurance and Financial Risk Management · Insurance, Mortality, Demography, Risk Management · Mathematics · Accounting