벤포드법칙과 회계부정; 감리지적기업을 중심으로

Benford’s law can be used to detect earnings management. The first accounting application of Benford’s law was by Carslaw(1988). Carslaw(1988) hypothesized that when net incomes are just below psychological boundaries, managers would tend to round these numbers up. A sign that such rounding-up was prevalent would be an excess of second digit zeros and a shortage of second digit nines. Carslaw(1988) used the expected second digit frequencies of Benford’s law to assess whether excess second digit zeros appeared in net incomes. Carslaw(1988) showed that there were more second digit zeros and fewer second digit nines than expected, thus suggesting that rounding-up of income did occur. Carslaw(1988) was followed by Thomas(1989). Thomas(1989) found the opposite effect for companies that reported losses. Thomas(1989) also found that earnings per share numbers were multiples of five cents more often than expected and had an ending digit nine less often than expected. Thomas(1989) suggested that earnings per share were rounded up and that net losses were rounded down. Skousen et al.(2004) found that Japanese firms also tend to round reported earnings. Kinnunen and Koskela(2003) examined the rounding phenomenon in eighteen countries, finding evidence of rounding behavior in most of the countries. Also, Benford’s law can be used to detect fraud. Nigrini(1996) used Benford’s law in a tax compliance context. Nigrini(1996) tabulated the first and second digit frequencies of interest received and total interest paid from a sample of tax returns. The data conformed to Benford’s law but there was a bias towards excess low digits for interest received and excess high digits for interest paid. Nigrini(1996) developed a distortion factor model that signals whether a tax field appears overstated or understated. The model assumes that authentic, unmanipulated data should conform to Benford’s law. The empirical results suggested that low income taxpayers were more likely to invent numbers on their tax return. In recent years, auditors have employed more sophisticated statistical techniques, such as digital analysis using Benford’s law, as part of their fraud detection processes. There is much empirical evidence suggesting that the frequencies of first and second digits of a data set that contains credible numbers will indeed correspond to a Benford’s law probability distribution(Nigrini 1996; Nigrini and Mittermaier 1997). Digital analysis based on Benford’s law is an audit technique that is applied to an entire population of transactional data. Benford’s law was introduced to the auditing literature in Nigrini and Mittermaier(1997), and researchers have since used these digit patterns to detect data anomalies by testing either the first, first-two, or last-two digit patterns of reported statistics or transactional data. Benford’s law routines are now included in software programs. This study estimated manipulation level of the accounting numbers in individual firms using Benford’s law and analysed whether this estimates can be used to detect financial statement fraud. For estimating manipulation level of the accounting numbers in individual firms, this study used the distortion factor model of Nigrini(1996). Nigrini(1996) developed the distortion factor model to estimate manipulation level in numbers with Benford’s law. The ABSDF, estimated absolute value using the distortion factor model, means the manipulation level of the accounting numbers in individual firms. Concretely, the ABSDF of firms sanctioned by Korean FSS(Financial Supervisory Service) compared to the ABSDF of matching firms. If the ABSDF of firms sanctioned by Korean FSS are higher than matching firms, ABSDF can be used to detect financial statement fraud. Findings for KSE and KOSDAQ listed firms from 2000 to 2010 are as follows. Results in mean difference showed that the ABSDF of firms sanctioned by FSS are higher than matching firms, and results in regression also showed that the ABSDF of firms sanctioned by FSS are higher than matching firms.