Charlie Munger: How To Identify Deceptive And Fraudulent Companies

Key Concepts

• Fraudulent Accounting Practices
• Complexity of Financial Statements
• Derivatives and Valuation Issues
• "Plug" Numbers in Accounting
• Accounting System Migrations and Discrepancies
• Competitive Frenzy and Out-of-Control Situations

Fraudulent and Complex Accounting

The discussion begins with an anecdote about a man attempting to sell a fire insurance company, stating, "It's like taking candy from babies." This individual specialized in insuring "concrete structures that are underwater," which the speaker immediately recognized as potentially fraudulent. The conversation then broadens to the inherent susceptibility of accounting to manipulation, particularly within financial institutions and the movie industry, where rapid write-offs of properties and construction in progress offer numerous avenues for deception.

The Deterioration of Financial Statement Usefulness

A key argument presented is that despite increased efforts to create more disclosures in accounting, present financial statements are not necessarily more useful, and in some cases, are less useful than those from 30-40 years ago. Charlie Munger elaborates that the financial statements of big banks are significantly harder to understand now due to their diversified operations, extensive footnotes, and what he terms "gobbledegook."

The Perils of Derivatives and Auditing Conflicts

The complexity extends to derivatives, with an example of a company owned that had 23,000 derivative contracts. Even with significant resources, including 10-20 math PhDs, understanding these contracts was impossible. This lack of comprehension cost approximately $400 million, even in a benign market. A critical issue highlighted is that the same auditing firm could audit two companies on opposite sides of a derivative transaction and attest to different values for the same contract. Charlie Munger recounts finding a mistake at Salomon Brothers on a derivative contract where both parties reported a large profit, blessed by their accountants, on the same contract. This situation is likened to a competitive frenzy where things spiral out of control.

The "Plug" Number at Salomon Brothers

A specific case study involves Warren Buffett becoming interim chairman and CEO of Salomon Brothers in 1991. He was informed of an item, approximately $180 million against a capital base of $4 billion, which was described as a "plug number." This number had been in place for 10 years, moving daily, as one entity (Fibre Merchant) operated on a trade date system while Salomon was on a settlement date system. Despite 10 years of auditing by Arthur Andersen and substantial auditing fees, the accountants could not resolve the discrepancy, leading them to "plug" the number daily. Buffett himself found it impossible to "unplug" without starting over.

Accounting System Migrations and Discrepancies

The concept of dealing with accounting discrepancies is further illustrated by an experience with a savings and loan. When changing accounting systems, a discrepancy arose that accountants could not fix. The solution adopted was to "let it run out" and then start over with a clean slate. This highlights that accounting is not a precise science and sometimes requires pragmatic, albeit unconventional, solutions.

Conclusion

The overarching takeaway is that accounting, despite its intended purpose of providing clarity, can become incredibly complex and opaque, especially in sophisticated financial markets and with large, diversified institutions. This complexity, coupled with competitive pressures and the inherent challenges of valuation, creates fertile ground for errors, manipulation, and a significant reduction in the usefulness of financial statements for investors and stakeholders. The anecdote about the Italian postal service, where "a few carloads" were discarded to improve efficiency, serves as a darkly humorous metaphor for how significant problems in accounting systems can sometimes be addressed by simply ignoring or eliminating problematic data.