First, the basic theory of financial fraud
(A) Triangle theory
The theory of fraud factors was first discussed by Mr. Lawrence Sawyer, who was called the father of internal audit by American academic circles. He pointed out that fraud is caused by three factors, that is, "a person must have three conditions if he wants to embezzle the employer's funds: abnormal needs, opportunities and rationality." This theory laid the foundation for the later development of fraud theory. Subsequently, Albert Lenchet, the founder of the American Institute of Certified Fraud Auditors and now the president of the American Accounting Association, put forward the triangle theory. In his view, the generation of enterprise fraud is composed of three elements: pressure, opportunity and excuse. Without any of the above elements, it is impossible to truly form corporate fraud. Therefore, to prevent the occurrence of enterprise fraud, not only as people usually think, but also by strengthening internal control, so as to eliminate the opportunity of enterprise fraudsters. In fact, we should also eliminate "pressure" or "excuses" to curb the occurrence of corporate fraud.
(b) outdated theories
GONE's fraud theory is the most popular in America. Enterprise fraud consists of four factors: G, O, N and E, which interact and are inseparable. No factor is more important than other factors, which together determine the degree of enterprise fraud risk. GONE consists of the first letters of four English words. Where g is greed, which means greed; O is opportunity, which means opportunity; N means need, which means need; E is exposure, which means exposure. The above four English words essentially express the four conditions of fraud, that is, when the fraudster is greedy and in urgent need of money, he will commit fraud as long as he has the opportunity and is not found afterwards. Therefore, the past tense theory conforms to the necessary and sufficient condition that when four factors work together and combine in a specific environment, the money, goods and rights of the deceived person will leave him.
(C) the theory of enterprise fraud risk factors
The theory of risk factors is developed by Bologna and others on the basis of Ghosn's fraud theory, and it is the most perfect theory about the risk factors that form enterprise fraud so far. According to this theory, the risk factors of enterprise fraud are composed of general risk factors and individual risk factors. First, the general risk factors refer to those factors that are mainly controlled by self-insurance organizations or entities, including: 1) the opportunity for potential corporate fraudsters to commit fraud; 2) the probability of being discovered when the company's fraud occurs; 3) The nature and degree of punishment for corporate fraudsters after discovering corporate fraud. Secondly, individual risk factors refer to those factors that vary from person to person and are beyond the control of organizations or groups, including moral quality and motivation.
Second, the research results of anti-fraud organizations
A large number of studies by scholars at home and abroad show that fraud is not a single act, and its occurrence depends on a series of conditions. As early as the 1930s, the United States began to study the anti-fraud strategies of enterprises, and in the 1970s, international anti-fraud theories began to appear. In economic fraud cases, the fraud in financial reports can not be ignored, which not only produces a lot of false accounting information, but also greatly increases the audit risk. The Treadway Committee (National Anti-Fraud Financial Reporting Committee) in the United States tries to solve the fraud problem by exposing fraud and detecting errors. The Committee described the definition of fraudulent financial report as "a deliberate or reckless act, whether it is a false report or an omission, the result is a major misleading financial report". In its epoch-making "Report 1987", Treadway Committee comprehensively expounded the anti-fraud prevention system of enterprises. It puts forward that any organizational entity can prevent enterprise fraud by establishing the following four lines of defense: high-level management concept; Internal control of business operation process; Internal audit; External independent audit. These control mechanisms complement each other, * * * together form an all-round and multi-level enterprise anti-fraud defense line, which can effectively check and deter enterprise fraud, that is, the anti-fraud four-level mechanism theory.
Third, build a diversified enterprise financial fraud audit countermeasures
The theory of financial fraud expounds the origin and breeding ground of fraud from different angles, and the corresponding anti-fraud results put forward measures to curb fraud from different levels, which not only provides theoretical basis for fraud audit, but also provides specific guidance for audit implementation. On this basis, this paper believes that fraud audit can detect the existence of financial fraud more accurately through quantitative testing and comprehensive results. A complete theoretical model composed of abnormal information, characteristic indexes and mathematical identification model is established, and the interaction mechanism of various factors is comprehensively analyzed. Diversified models can expand the ideas and means of enterprise fraud audit.
(A) choose financial indicators as the cornerstone of the study.
The foundation moistens the wind and the moon is dizzy and rainy. The research shows that the relationship between financial indicators and financial fraud means is universal, and the manipulation means can be exposed by comparing the cross-test relationship between financial indicators. If we pay enough attention to these financial indicators that reflect profit manipulation, we can identify fraudulent means to a great extent.
(2) Build a combination recognition model.
Establishing mathematical models can often effectively identify financial fraud. In order to improve the accuracy of the identification model, it is necessary to adopt multi-level and multi-angle forecasting methods, including not only the anomalies of financial figures, but also the indicators reflecting the factors inducing corporate fraud and the lack of control environment. So as to explore the relationship between these data, use database technology to screen out useful data from the sample database, eliminate useless information, establish Probit discriminant model, linear probability model and Logistic regression model, and combine these different recognition models appropriately to get better combined recognition results than any independent recognition. Considering the reliability and risk of identification, it is not feasible to identify complex financial fraud activities only by using a single identification model, and the information used by a single different identification model will not be the same. Almost any model will contain some useful independent information, so giving up an identification method may lead to the underutilization of valuable economic information resources. In order to solve this limitation, a combined identification model is proposed. It synthesizes the identification results of various fraud identification methods and obtains a combined identification model that may be better than any independent identification model.
(3) combination of characteristic signals
The transience and unpredictability of the market make the information authenticity risk exist objectively. The abnormality of indicators can play a warning role, but it is still thin to identify the traces of financial fraud only through data analysis. Cooper &; The 29 red flags summarized by Lybrand accounting firm as the possible basis for financial reporting fraud have given us good enlightenment. When we study China's financial fraud, we should not stay at the superficial phenomenon, but start with the root of financial fraud and pay attention to the factors that lead to fraud. Through normative research, we collect, summarize and sort out the techniques and means of accounting experts to identify financial fraud, and build an incomparable signal set as a weather vane to warn the existence of false information.
(D) the establishment of a comprehensive index evaluation system
The identification model intuitively describes the authenticity of enterprise accounting information, but a single result cannot reproduce the situation of enterprise accounting information, and pure qualitative analysis often lacks sufficient evidence. Therefore, it is undoubtedly a good method to establish a set of index system to identify financial fraud by using fuzzy mathematics, and to combine the two to make up for their respective shortcomings. The index system can adopt the method of fuzzy comprehensive evaluation, and one of the important tasks is to determine the single factor evaluation set of influencing factors, which is often very difficult. On the one hand, when there are many influencing factors, it is a huge workload to determine each parameter one by one; On the other hand, the parameters given by experts according to their own experience often need to be adjusted according to the actual situation. When adjusting data, it is difficult to operate because a large number of data are interrelated. Therefore, the combination of fuzzy mathematics and neural network is used to set weights, and the powerful data processing and learning ability of neural network is skillfully used to form a fuzzy neural network, which can not only explain qualitatively, but also process quantitatively, and is widely used in data processing and fuzzy neural network of non-qualitative systems. In the process of learning a large number of examples, the membership degree and weight are constantly revised, which greatly improves the recognition success rate of the index system.
In the market economy, people can't completely eliminate the asymmetry of accounting information and accounting uncertainty through laws and regulations. As long as there is principal-agent relationship and different stakeholders, it is difficult to eliminate the fraud motivation of enterprise management, and financial fraud will continue to exist. In other words, it is impossible for Pareto or even Pareto to improve accounting information under pure market conditions. This requires in-depth study of the development law and identification methods of financial fraud in the field of fraud audit, and provides empirical support for fraud identification, real-time monitoring and comprehensive management.