The scientific background of zapliance can hardly be denied. After all, the foundation of our company can be traced back to the University of Hamburg, where the following story unfolded a few years ago…
It was a humid Friday in the summer of 2014. Prof. Dr. Nick Gehrke and Prof. Dr. Frank Rump were attending the defense of a dissertation that was being held at the University of Hamburg as part of the research project called “Virtual Accounting Worlds”. The project was the cornerstone for the development of the Financial Process Mining Algorithm that is today put to work in zap Audit.
After the defense, Nick and Frank discussed a problem with the algorithm on the whiteboard. But as so often when two strong opinions collide, no agreement could be reached, and so they decided to make a bet with each other. The task was clear:
To minimize the runtime which at that time was 12 days.
In the end, the winner would be the one whose algorithm was faster.
Another bet was also made between business informatics (represented by Prof. Dr. Nick Gehrke) and computer science/mathematics (Prof. Dr. Frank Rump). Over the weekend from Friday to Monday, quite a few liters of coffee were consumed, and various ideas were critically examined and then discarded. On Monday, both optimizations were tested at the same time. The results were unambiguous:
Original runtime: 12 days
Prof. Dr. Nick Gehrke’s algorithm took three days to complete the test.
Prof. Dr. Frank Rump could only smile when his algorithm completed the test in just 12 minutes.
Why am I telling you this?
– Because this story clearly shows what effects science can have on the development of solutions to a problem and how it can provide stimuli for new approaches.
For this reason, we have decided to launch a series entitled “Audit in science”. At irregular intervals, we will, therefore, present 2-3 scientific publications from the field of auditing. There will be something for everyone and the contributions will certainly provide food for thought. At the end of the articles, we look forward to hearing your feedback and would like to use the results to take a more in-depth look at one or the other of the topics, if necessary.
I will provide a brief introduction to each of the publications concerned, at the end of which you find a download link to the publication in question.
So, without further ado, let’s get started.
Accounting fraud in the 21st century
The first publication is a review of literature on the subject of “Research Topics in Accounting Fraud in the 21st Century: A State of the Art” and examines, among other things, the importance of responsible corporate management and good accounting practices:
Companies play a role in society that goes beyond mere economic interest. Their contribution to social development and to the sustainability of the territory where they are located seems unquestionable. However, after the great financial scandals of companies such as ENRON, WorldCom or AHOLD, interest groups require accurate and transparent financial information. The development of more demanding financial reporting standards seems, however, not to have been up to scratch, since accounting fraud continues to be detected all over the world. The search, therefore, for possible causes that may induce companies to act unethically was the main motivation behind this research. To do this, a review of the literature in high-impact journals that have dealt with accounting fraud, covering the main lines of research, was carried out. The findings of the literature review highlight the importance of responsible corporate governance and good accounting practices, as well as the importance of certain psychological characteristics of managers and employees as enhancers of the lack of ethics. It is clear that the social cost of accounting fraud should be minimized, and governments should develop specific policies that combine responsible corporate governance in companies with the sustainability of their environment.
The detection of tax fraud with neural networks
The second publication comes from the field of tax fraud. For many companies, this is a sensitive issue and a bit of a “problem child”, but its relevance will never diminish. Although this publication deals with the percentage of taxpayers who have a propensity to commit tax fraud, the approach is transferable to some extent to other corporate use cases. Machine learning methods are used to investigate personal income tax returns:
The goal of the present research is to contribute to the detection of tax fraud concerning personal income tax returns (IRPF, in Spanish) filed in Spain, through the use of Machine Learning advanced predictive tools, by applying Multilayer Perceptron neural network (MLP) models. The possibilities springing from these techniques have been applied to a broad range of personal income return data supplied by the Institute of Fiscal Studies (IEF). The use of the neural networks enabled taxpayer segmentation as well as the calculation of the probability concerning an individual taxpayer’s propensity to attempt to evade taxes. The results showed that the selected model has an efficiency rate of 84.3%, implying an improvement in other models utilized in tax fraud detection. The proposal can be generalized to quantify an individual’s propensity to commit fraud with regards to different kinds of taxes. These models will support tax offices to help them arrive at the best decisions regarding action plans to combat tax fraud.
What do you think?
Shall we go into one of the topics in more detail and, for example, present some MLP models? The choice is yours: