Research On Time-series Fraud Characteristics Of Manipulating Profits Companies

At the beginning of the21st century, financial fraud problems seem to have become the focus of global issues, and accounting scandals increasingly destroy the public confidence. In October2013, the China Securities Regulatory Commission carried out a biggest-ever financial check which was faced to the company prepared to list. This measure made268companies (accounted for more than30%) give up the listing application plan. This means that even these so called good enterprises, there are still financial problems that we can’t ignore, and even financial frauds. The increasingly serious fraud trend not only brings huge losses to our country and organizations, but also makes the investors and the social public lose confidence for the capital market.To implement effective regulation for financial fraud of listed company, we must analyze the fraud motivation and means from the beginning and explore significant anomalous characteristics between the fraudulent companies and the normal companies, which will improve the regulation efficiency. This paper puts the profit manipulation which is the most typical and serious type of fraud as the breakthrough point to clarify the difference between the fraudulent companies and the normal ones. On the basis of existing research results, this paper introduces the concept of time-series into research, analyzes the fraud data of consecutive four years, and puts the financial statement subjects which have been ignored by many scholars into the research scope. This may achieve the goal of identifying financial fraud characteristics from the start.This article selects fraudulent companies from the time when criminal records is exposed to us to Dec.31st,2013. All the companies are A-share listed in Shanghai and Shenzhen of China. First of all, the financial fraud characteristics of industry, year and type are analyzed on the whole. Secondly, profit manipulation company samples are picked out, as well as the control samples of the same number. Through the two methods of independent sample T test and the Mann-Whitney U test, this paper desires to test if there is significant difference between two group sample data of fraudulent year and the three years before fraud. Financial statement subjects and financial indexes are two important aspects to make analysis and explanation. The correlation between them is also explored in this paper, which gets meaningful conclusions. At last, on the foundation of the test result, a logistic regression model based on time-series is set up successfully, which will provide a new fraud recognition reference for regulatory agencies.