Abstract  

MCA Technology Solutions Private Limited was established in 2015 in Bangalore with an objective to integrate analytics and technology with business. MCA Technology Solutions helped its clients in areas such as customer intelligence, forecasting, optimization, risk assessment, web analytics, and text mining and cloud solutions. Risk assessment vertical at MCA technology solutions focused on problems such as fraud detection and credit scoring. Sachin Kumar, Director at MCA Technology Solutions, Bangalore was approached by one his clients, a commercial bank, to assist them in detecting earnings manipulators among the bank’s customers. The bank provided business loans to small and medium enterprises and the value of loan ranged from INR 10 million to 500 million. The bank suspected that its customers may be involved in earnings manipulations to increase their chance of securing a loan. 

Saurabh Rishi, the chief data scientist at MCA Technologies was assigned the task of developing a use case for predicting earnings manipulations. He was aware of models such as Beneish model that was used for predicting earnings manipulations; however, he was not sure of its performance, especially in the Indian context.  Saurabh decided to develop his own model for predicting earnings manipulations using data downloaded from the Prowess database maintained by the Centre of Monitoring Indian Economy (CMIE). Daniel received information related to earning manipulators from Securities Exchange Board of India (SEBI) and the Lexis Nexis database. Data on more than 1,000 companies was collected to develop the model.

Learning Objective  

The case may be used in Business Analytics and Fraud Analytics courses of MBA or Executive MBA programs to teach fraud/earnings manipulations prediction. The learning objectives are as follows.

1. Understand difference between earnings management, earnings manipulations, and accounting fraud.
2. Learn how predictive analytics techniques can be used to predict earnings manipulations by firms.
3. Learn concepts such as Youden’s index and study how it can be used for calculating the optimal classification cut-off probability.