scholarly journals Detecting Financial Statement Fraud with Interpretable Machine Learning

2021 ◽  
Author(s):  
Zhongzhu Liu ◽  
Rongguang Ye ◽  
Rongye Ye
Keyword(s):  
Adaptive Lasso ◽  
Financial Statement ◽  
Financial Fraud ◽  
Integrated Models ◽  
Financial Statement Fraud ◽  
Dependence Function ◽  
Interpretable Machine Learning ◽  
Auc Value ◽  
Ks Values ◽  
Partial Dependence

Abstract In this study, we explored a stable and explainable model in the detection of financial fraud. To effectively handle imbalanced datasets, we selected the Smote oversampling algorithm with the highest AUC value and compared it with Borderline Smote and ADASYN algorithms. Using the MCB method, we found that the Adaptive Lasso algorithm had higher stability than SCAD, MCP, Stepwise, and SQRT Lasso algorithms. Moreover, the AUC value was improved by WoE encoding and IV value testing of the features. Finally, we ranked the fraud factors based on the importance of the features, and the partial dependence function was used to make the model interpretable. By comparing the AUC and KS values, the integrated models XGBoost, LightGBM, and RF showed better ability to identify financial fraud compared with traditional models such as SVM and LR.

Fintech in financial reporting and audit for fraud prevention and safeguarding equity investments

2020 ◽  
Vol ahead-of-print (ahead-of-print) ◽  
Author(s):  
Paulina Roszkowska
Keyword(s):  
Financial Reporting ◽  
Financial Statements ◽  
Financial Statement ◽  
Financial Fraud ◽  
Financial Statement Fraud ◽  
Content Type ◽  
Financial Scandals ◽  
External Auditor ◽  
Equity Investments

Purpose The purpose of this paper is to explore the audit-related causes of financial scandals and advice on how emerging technologies can provide solutions thereto. Specifically, this study seeks to look at the facilitators of financial statement fraud and explain specific fintech advancements that contribute to financial information reliability for equity investments. Design/methodology/approach The study uses the case studies of Enron and Arthur Andersen to document the evidence of audit-related issues in historical financial scandals. Then, a comprehensive and interdisciplinary literature review at the intersection of business, accounting and engineering, provides a foundation to propose technology advancements that can solve identified problems in accounting and auditing. Findings The findings show that blockchain, internet of things, smart contracts and artificial intelligence solutions have different functionality and can effectively solve various financial reporting and audit-related problems. Jointly, they have a strong potential to enhance the reliability of the information in financial statements and generally change how companies operate. Practical implications The proposed and explained technology advancements should be of interest to all publicly listed companies and investors, as they can help safeguard equity investments, thus build investors’ trust towards the company. Social implications Aside from implications for capital markets participants, the study findings can materially benefit various stakeholder groups, the broader company environment and the economy. Originality/value This is the first paper that seeks solutions to financial fraud and audit-related financial scandals in technology and not in implementing yet another regulation. Given the recent technology advancements, the study findings provide insights into how the role of an external auditor might evolve in the future.

scholarly journals ANALISIS FRAUD DIAMOND DALAM MENDETEKSI FINANCIAL STATEMENT FRAUD: STUDI EMPIRIS PADA PERUSAHAAN MANUFAKTUR YANG TERDAFTAR DI BURSA EFEK INDONESIA (BEI) TAHUN 2016 – 2018

2019 ◽  
pp. 2
Author(s):  
Ika Cipta Suryani
Keyword(s):  
Pressure Change ◽  
Financial Statement ◽  
Financial Fraud ◽  
Financial Statement Fraud ◽  
Target Pressure

<p>Penelitian ini dilakukan untuk menganalisis pengaruh variabel Fraud Diamond terhadap Financial Fraud Statement. Sampel yang digunakan dalam penelitian ini adalah 28 perusahaan manufaktur yang terdaftar di Bursa Efek Indonesia selama periode 2016-2018. Jenis data yang digunakan adalah data sekunder, berupa laporan tahunan perusahaan yang terdaftar di Bursa Efek manufaktur selama periode 2016-2018. Pengujian hipotesis dilakukan dengan menggunakan regresi linier berganda dengan perangkat lunak Eviews 8. Hasil penelitian menunjukkan bahwa variabel Financial Target (Pressure), Change in Auditor (Rationalization), dan Change in Director (Capability) berpengaruh secara signifikan terhadap Financial statement, sedangkan Ineffective Monitoring (Opportunity) tidak berpengeruh terhadap Financial Statement Fraud</p>

Strain, Differential Association, and Coercion: Insights from the Criminology Literature on Causes of Accountant's Misconduct

2008 ◽  
Vol 8 (1) ◽  
pp. 1-20 ◽  
Author(s):  
James J. Donegan ◽  
Michele W. Ganon
Keyword(s):  
Criminal Behavior ◽  
White Collar Crime ◽  
Policy Implications ◽  
Financial Statement ◽  
Differential Association ◽  
Financial Fraud ◽  
Certified Public Accountants ◽  
Fraud Triangle ◽  
Financial Statement Fraud ◽  
Coercion Theory

This paper introduces to the accounting literature two prominent criminology theories, strain and differential association, as possible explanations for criminal behavior by accountants and applies a recent integration of the two, coercion theory, to three recent financial statement frauds. We argue that understanding and preventing fraudulent accounting can be furthered by placing the phenomenon within the context of criminology research, which supports both individual and group-level explanations for white-collar crime. We also suggest that the American Institute of Certified Public Accountants (AICPA) moved too quickly in adopting Cressey's fraud triangle as the explanatory model for financial fraud in Statement on Auditing Standards (SAS) No. 99. Our analysis, although exploratory in nature, suggests that examining financial statement fraud through the lens provided by criminology theory may provide new insights into its causes as well as tools for detection and prevention. We conclude with a discussion of policy implications.

Feature Selection for Dimension Reduction of Financial Data for Detection of Financial Statement Frauds in Context to Indian Companies

2020 ◽  
pp. 097215092092866
Author(s):  
Sonika Gupta ◽  
Sushil Kumar Mehta
Keyword(s):  
Feature Selection ◽  
Selection Process ◽  
Fraud Detection ◽  
Original Data ◽  
Financial Data ◽  
Financial Statement ◽  
Financial Fraud ◽  
Financial Statement Fraud ◽  
Financial Variables ◽  
Input Dataset

The financial fraud detection problem involves analysis of the large financial datasets. Financial statement fraud detection process is concentrated on two major aspects: first, identification of the financial variables and ratios, also termed as features. Second, applying the data mining methods to classify the organizations into two broad categories: fraudulent and non-fraudulent organizations. If the input dataset contains large number of irrelevant and correlated features, the computational load of the machine learning technique increases and the effectiveness of the classification outcomes decreases. The feature selection process selects a subset of most significant attributes or variables that can be the representative of original data. This selected subset can help in learning the pattern in data at much less time and with accuracy, in order to produce useful information for decision-making. This article briefly states the methods applied in the prior studies for selecting the features for financial statement fraud detection. This article also presents an approach to feature selection using correlation-based filter selection methods in which feature selection is performed based on ensemble model, and tests the outcome of the approach by applying the mean ratio analysis on financial data of Indian companies.

scholarly journals DETERMINANT OF FRAUD PENTAGON IN DETECTING FINANCE OF FINANCIAL STATEMENTS

2020 ◽  
Vol 4 (1) ◽  
pp. 32-41
Author(s):  
Muhammad Bagus ◽  
Noviansyah Rizal ◽  
Siwidyah Desi Lastianti
Keyword(s):  
Internal Pressure ◽  
Stock Exchange ◽  
Financial Statements ◽  
Financial Statement ◽  
Financial Fraud ◽  
Test Results ◽  
Financial Statement Fraud ◽  
Score Model ◽  
Fraudulent Financial Statements ◽  
Industry Conditions

This study aims to determine the Pentagon Determinant Fraud in detecting fraudulent financial statements. Fraudulent financial statements are proxied by the Fraud Score Model. Whereas the pressure factor is proxied by insisting from within, for the opportunity factor proxied by industry conditions, the rationalization factor is proxied by the ratio of total accruals, the competency factor is proxied by the change of directors and arrogance is proxied by the duality of quality positions at the CEO. The population in this study amounted to 100 companies incorporated in the compass index 100 contained in the Indonesia Stock Exchange and for the sample of the study were 35 companies belonging to the compass index 100 contained in the Indonesia Stock Exchange, which was selected using the purposive sampling method for the 2017-2018 period. Data were analyzed using multiple linear regression. Based on the test results, it was concluded that the pentagon fraud component included internal pressure (LEV), industry conditions (INVENTORY), rationalization (TATA) influencing financial statement fraud while competence (DCHANGE) and arrogance (DCD) had no effect on financial fraud statement. This proves that internal pressure (LEV), industry conditions (INVENTORY), and rationalization (TATA) can be used to detect fraud in financial statements.

Real earnings management and financial statement fraud: evidence from Malaysia

2018 ◽  
Vol 26 (4) ◽  
pp. 508-526 ◽  
Author(s):  
Noorul Azwin binti Md Nasir ◽  
Muhammad Jahangir Ali ◽  
Rushdi M.R. Razzaque ◽  
Kamran Ahmed
Keyword(s):  
Earnings Management ◽  
Cash Flow ◽  
Design Methodology ◽  
Production Costs ◽  
Real Earnings Management ◽  
Financial Statement ◽  
Financial Fraud ◽  
Financial Statement Fraud ◽  
Content Type

Purpose We examine whether the fraud firms are engaged in real earnings management and accrual earnings management prior to the fraud year in the Malaysian context. Design/methodology/approach Our sample comprises of 65 financial statement fraud and 65 non-fraud firms over a period of eight years from 2001 to 2008. Findings Using the abnormal cash flow from operations (CFO) and abnormal production costs as the proxies for real earnings management, we find that financial statement fraud firms engage in manipulating production costs during preceding two years of the fraud event. However, our results show that financial fraud firms engage in manipulating CFO prior to the fraud event. Additionally, we find that financial statement fraud firms prefer to manipulate earnings using accruals relative to real earnings prior to the fraud year. Originality/value Our results demonstrate that real earnings management is more aggressive in financial statement fraud firms compared to the non-fraud firms in the four years prior to fraud.

Detecting the probability of financial fraud due to earnings manipulation in companies listed in Athens Stock Exchange Market

2021 ◽  
Vol ahead-of-print (ahead-of-print) ◽  
Author(s):  
Andreas Maniatis
Keyword(s):  
Financial Services ◽  
Stock Exchange ◽  
Financial Statement ◽  
Financial Fraud ◽  
Earnings Manipulation ◽  
Exchange Market ◽  
Financial Statement Fraud ◽  
Content Type ◽  
General Index ◽  
Athens Stock Exchange

Purpose The aim of this paper is to detect whether there are companies listed in the general index of Athens Stock Exchange Market that possibly conduct earnings manipulation during 2017–2018. Design/methodology/approach The paper is based upon the Beneish model (M-score), which consists of eight variables to examine the probability of financial statement fraud related to earnings manipulation for 40 companies listed in the Athens Stock Exchange Market. Any company with an M-score −2.22 or above is likely to be a manipulator whereas any company that scores −2.22 or less is unlikely to conduct earnings manipulation. Findings After calculating the M-score for each company, it was found that 33 (out of 40) companies had M-score values lower than −2.22. Therefore, 82.5% of the sample is considered rather unlikely to conduct earnings manipulation whereas 17.5% of the companies listed in the general index of Athens Stock Exchange Market is likely to manipulate its earnings. Research limitations/implications In this paper, all institutions related to financial services were left out of the sample because of the fact that M-score cannot provide reliable results when applied on similar companies. Originality/value Beneish model offers a probability of financial fraud and can be therefore used as a supplementary test for auditors, fraud examiners or even national regulators such as the Hellenic Accounting and Auditing Standards Oversight Board or the Hellenic Capital Market Commission. The results of this paper can contribute to the literature concerning financial fraud in Greece during 2017–2018 because no relevant recent researches have been published yet.

scholarly journals CORRELATION OF FINANCIAL STATEMENT COMPONENTS IN DETECTING FINANCIAL FRAUD

2016 ◽  
Vol 1 (2) ◽  
pp. 275
Author(s):  
Handayani Handayani ◽  
Tarjo Tarjo ◽  
Yuni Rimawati
Keyword(s):  
Capital Market ◽  
Sampling Technique ◽  
Financial Statements ◽  
Cash Flows ◽  
Financial Statement ◽  
Financial Fraud ◽  
Spearman Correlation ◽  
Red Flags ◽  
Financial Statement Fraud ◽  
Spearman Correlation Test

The purpose of this study is to determine the absence of correlation offinancial statement components as red flags in detecting financial statement fraud. The sampling in this study is done using purposive sampling technique. There are two categories of companies used as the study sample:fraud companies and non-fraud companies. Fraud companies are the companies that get sanction from Capital Market Supervisory Agency (Bapepam) andFinancial Services Authority (OJK) period 2000-2014, while non-fraud companies are selected with the criteria: having equivalent assets, engaging in the same industry, and using the same financial statements as the financial statements used by the fraud companies. The total samples of this study are 122 companies consisting of 61 fraud companies and 61 non-fraud companies. Spearman correlation test is used to answer the research hypothesis.The conclusions of this study are (1) the absence of correlation between cash flows andearnings can be usedas red flags to detect fraud, (2) the existence of correlation between receivables and revenues cannot be used as red flags to detect fraud, (3) the existence of correlation between allowances for uncollectible accounts and receivable cannot be used as red flags to detect fraud.

Pengaruh Faktor Demografi Dan Iklim Organisasi Terhadap Niat Whistleblowing Internal

2017 ◽  
Vol 4 (1) ◽  
pp. 12-22
Author(s):  
Priyastiwi Priyastiwi
Keyword(s):  
Data Analysis ◽  
Data Collection ◽  
Organizational Climate ◽  
Demographic Factors ◽  
Financial Statement ◽  
Research Use ◽  
Analysis Method ◽  
Financial Statement Fraud ◽  
Internal Whistleblowing ◽  
Collection Methods

This study aimed to examine the effect of demographic factors and organizational climate on the intention of internal whistleblowing. The sample was an accountant who had worked as an auditor. Data collection methods using questionnaires with financial statement fraud case scenarios. This research use ANOVA data analysis method to examine demographic factors include age, gender, and experience, as well as organizational climate on the intention of internal whistleblowing. The results showed only the experience of demographic factors that influence internal whistleblowing. Besides internal whistlelowing also influenced by organizational climate in the company.Keyword: Demographics, Organizational Climate, Whistleblowing

Sign in / Sign up

Export Citation Format

Share Document