A model recommendation on the determination of financial statement manipulation risk: BİST application

The most basic tools and primary sources of information about the economic situation of businesses are the financial statements. Information obtained from financial statements are the most important elements used by decision makers in decision making. The manipulation of financial statements, which can be expressed simply as the presentation of financial information as different from that of reality, places great burdens on economic circles. The aim of this work is to develop an effective model that can be used to detect the manipulation of financial statement. In this framework, the financial statement data of 403 companies traded in BIST were subjected to Benford analysis and searched the manipulation / fraud signal in financial information. The Benford Digit Score (BBS), which can be an alternative measurement standard with new critical values in the evaluation of digit test results, has been developed in the study of evidence that conforms to the Benford's law. In the analysis of the differences according to the BBS value, it has been determined that the size of the audit firm and the requirement of independent audit have a significant effect on the accuracy of the financial statements. There were also significant differences between the BIST risk groupings and sectors in terms of financial statement accuracy. In the study a new model is proposed as an alternative to the mixed models used in the determination of manipulation of the financial statements over 184 BIST companies according to the transactions in the real sector. In the literature, the 38 financial ratios selected from financial ratios frequently used in manipulation detection were tested by logistic regression analysis and the 7 rate model with the highest manipulative detection power was included. As a result, it was observed that Benford's analysis was effective in measuring the accuracy of the financial statements, but that the comparative interpretation of the calculated BBS values would increase the reliability. It has been found that the newly developed logistic regression model correctly predicted manipulated companies by 16.7%, with or without manipulation as 66.6% on average. Compared to other mixed models, level of significance and clarity are higher.