Marcellus’s 3-Level Check To Detect Accounting Frauds

In this section, we will see how the author explained the process of detecting accounting fraud in his book. 

Evaluating accounting quality and corporate governance is the cornerstone of Marcellus's investment process. Their research covers this evaluation in three ways: Level 1 checks, where they use their quantitative forensic screens to rank companies on their accounting quality. Their quantitative screens reject nearly 40% of companies from broad-based stock exchange indices that they analyse. The companies which make it through their quantitative forensic screens are the ones on which they conduct in-depth research and diligence. These companies are subject to two different levels of forensic checks: Level 2 checks, where they conduct an in-depth reading of the financial statements to identify some of the traits mentioned in the preceding section; and finally, Level 3 checks, where the author also tries to gauge promoter/management integrity by doing primary data checks on the companies, wherein the author speaks with former employees, customers and suppliers of the company.

Level 1 Checks: Marcellus's Clean Accounting Model

The first line of defence is the author's proprietary forensic accounting model. They have developed a model using a set of twelve ratios that helps us grade companies on their accounting quality. The selection of these ratios has been inspired by Howard M. Schilt's legendary book on forensic accounting, Financial Shenanigans. 

This book draws upon case studies of accounting frauds (involving well-known frauds like Enron and WorldCom and numerous lesser-known cases of accounting trickery) and, from the lessons learnt, creates techniques for detecting frauds in financial statements.

Their twelve forensic accounting ratios cover checks across the critical components of financial statements, including the profit and loss account (for revenue or earnings manipulation), balance sheet (for false representation of assets and liabilities), cash pilferage and audit quality. Some of the key ratios and their rationale are shown below.

The Forensic Framework has proven to be an effective predictive tool

There is a strong correlation between accounting quality, as indicated by their forensic model, and shareholder returns. 

Level 2 Check: Further Accounting and Corporate Governance checks

These checks are over and above the forensic screens and are usually employed during their deep-dive diligence on individual companies. Their forensic framework helps us weed out companies with dubious accounting quality. It also helps us identify the key accounting red flags for a company. However, their forensic model may not be able to pick up several qualitative aspects of accounting and corporate governance due to a lack of uniformity in data across companies or where there is subjective judgment involved. Such areas can only be evaluated through a deep dive into historical financial statements and primary data checks around management integrity. They have developed the following checklist for deeper accounting and corporate governance checks beyond the forensic model.

This checklist should form an essential part of qualitative assessment of any stock.

Level 3 Checks: Extensive Diligence through Primary Data Checks

The integrity of a company's promoters and management is assessed as the next level of accounting quality checks through detailed discussions with our primary data sources. The author seeks confirmation of the picture painted by the company's financials from primary data sources familiar with the company and who have worked with or dealt with it. Consumers of the company's products, channel partners, raw material suppliers, competitors, and so on are examples of such sources. This is especially important in the case of small caps, which have a shorter operating history and a high propensity to fudge accounts. Similarly, visiting the company's branches, factories, or even a plant under construction helps build confidence in the company's financials.