Hidden assets: concealment, tracing and recovery

A whistleblower tells the company that its procurement manager owns three houses registered in the names of his cousins, a shell company in another country and a crypto wallet. On paper, the manager has a modest salary and no property. If the company wins its case, the court will order him to pay back what he took, but only assets that can be found can be used to pay.

The forensic accountant's task is to find them, link them to the person and show that the money came from the fraud.

This topic covers what hidden assets are, how and why people hide them, the types of hidden assets, the methods used to find them, forensic techniques, data analytics and machine learning in the search, and their benefits and limits. The previous topic covered computer forensics and fraud investigation, and the next covers professional negligence.

Key terms in plain English

Syllabus check

This topic is in syllabus area D, forensic accounting services (15% of marks). It covers:

• hidden assets: methods of concealment and detection
• forensic techniques and data analytics in tracing assets
• machine learning techniques

CAMA 2020 and hidden assets

The Companies and Allied Matters Act 2020 (CAMA 2020) is the company law of Nigeria. The sections below apply to this topic.

The rules in more depth

What hidden assets are

Hidden assets, in forensic accounting, are assets of value that are deliberately concealed to prevent their detection in legal proceedings, in an investigation or during the resolution of a dispute over ownership.

The reasons for hiding them vary: to avoid paying a court judgment, to keep the proceeds of fraud or corruption, to avoid tax, to reduce a share in a divorce or in a shareholder or partnership dispute, to escape creditors, or to hide money laundered from crime.

How assets are hidden

• False reporting: deliberately allowing material misstatement in financial statements, for example by understating assets or overstating liabilities.
• Nominees: holding assets in the names of other people or in nominee accounts, including cryptocurrency held in accounts under other names.
• Shell companies: creating companies with no real business, owned by nominees, to hold assets.
• Offshore accounts and tax havens: using bank accounts abroad or places with strict secrecy to conceal assets.

Other methods are transfers to relatives and friends before a case begins, delayed receipt of income, overpaying tax or credit cards to get a refund later, buying valuable art, jewellery or precious metals, and paying fake debts to friendly creditors.

Types of hidden assets

The process of finding hidden assets

Finding hidden assets combines several approaches:

1. Research. Gather all relevant information from financial records, business transactions and personal relationships of the person or organisation involved. Review transaction documents and financial statements to find any potential investments.
2. Follow-up and analysis.
   • Asset tracing: through the statement of financial position if an organisation is involved, and through bank statements.
   • Public records: deeds, court documents and survey plans for land.
   • Third-party search: reviewing third parties such as accountants, attorneys and other experts who may know of assets.
3. Investigation techniques. Interview and interrogate all persons who know about the asset; and use surveillance to gather information about assets.
4. Forensic analysis. Use digital or financial analysis to identify and follow assets.
5. Machine learning. Use algorithms to find potential assets or transactions.
6. Legal tools. Use court orders to reach undisclosed assets.

Techniques for tracing hidden assets

• Data analytics: analysing data and transactions to find patterns and anomalies.
• Documentation: reviewing documents such as bank statements and tax returns to find discrepancies.
• Network analysis: analysing relationships and links between individuals and entities to find possible hidden assets.
• Interviews and interrogation: speaking with everyone involved to gather information that may lead to the assets.
• Digital analysis: analysing digital data, such as e-mails and computer logs, for evidence of hidden assets.

In practice the investigator also uses the net worth method: compare the person's assets at the start and end of a period, add spending and subtract known lawful income. If the result is more than known income, the difference may come from hidden sources. The source and application of funds method does the same by listing every inflow and outflow.

Registry searches at the Corporate Affairs Commission for directorships and shareholdings, the land registry for property and vehicle registration records are used to find assets in the Nigerian setting, and mutual legal assistance or foreign registries for assets abroad.

Using data analytics to find hidden assets

Data analytics plays a key role. It finds hidden assets by:

• identifying patterns and anomalies in financial data that may show hidden assets;
• analysing large data sets, including transactions, account statements and tax returns, to show where assets may be hidden; and
• detecting unusual transactions in account statements and tax returns.

Analytics techniques include machine learning, which uses algorithms to find patterns and anomalies; cluster analysis, which groups similar transactions or account activities; predictive modelling, which estimates the likelihood of hidden assets from past data; and network analysis, which studies relationships between people and entities.

Data sources include financial statements (income statement, financial position and cash flow statements), bank statements, which are valuable in money laundering cases, tax returns, which show income, expenses and assets, and public records such as property records and court filings.

Benefits: analytics improves accuracy by finding patterns that manual analysis may miss; improves efficiency by automating the analysis of large data sets; and gives insight into relationships between individuals and entities. Challenges: the quality of the data; the specialised skills needed; and time, because the investigation must be done as quickly as possible to prevent the assets being dissipated.

Machine learning techniques

Machine learning analyses large data sets to find patterns that may show hidden assets.

• Deep learning: algorithms that analyse complex data to find patterns and anomalies. An algorithm is a clearly defined set of step-by-step instructions that takes input, works on it and gives output to solve a problem.
• Supervised learning: algorithms such as decision trees and random forests, trained on labelled data sets, to find patterns that indicate hidden assets. A random forest combines many decision trees to improve the accuracy and strength of its predictions.
• Unsupervised learning: algorithms such as clustering and dimensionality reduction, which find patterns without labels.

Applications: financial statement analysis, to find inconsistencies and anomalies; transaction analysis of bank statements and card records, to find suspicious activity; and network analysis of data such as social media, e-mail communications, video conferences and messaging conversations, to find relationships and patterns.

Limitations:

• Data quality: poor data reduces the accuracy and reliability of models.
• Complexity: it can be hard to interpret what the model produces, so the reason for a decision may not be fully understood.
• Overfitting: a model that is too specialised to its training data learns the noise, which is random fluctuation or error in the data, instead of the real pattern, and fails to generalise to new data. Training data is the set of examples used to train the machine, for example 1,000 invoice values used to teach it to list receivables of ₦10,000 and above.

Recovery of the assets

Finding the asset is only part of the task. The client's lawyers may need to obtain court orders to freeze the asset so that it cannot be sold or moved, to require disclosure of assets, or to appoint a receiver. The forensic accountant supports this with a clear report that links the asset to the person, shows how it was funded and states its value.

The report shows the tracing step by step, so that the court can follow the money from the fraud to the asset. If the asset is held by a third party who gave nothing for it, such as a relative, the court may treat the transfer as a gift made to defeat creditors and set it aside.

Ethical and legal limits apply. The investigator must not obtain information illegally, such as by hacking an account or bribing a bank official, and must respect privacy and data protection law. Surveillance must be lawful and proportionate. The evidence must be gathered in a form that the court will accept.

Worked examples

1. The increase is ₦85 million, but lawful savings were only ₦17 million.
2. ₦68 million is unexplained and may come from hidden sources.

1. The cousins are nominees.
2. The payments link the houses to the manager, and the court may order the properties to be treated as his.

1. The company has no activity and its only income is from the supplier.
2. Network analysis links the supplier, the shell company and the manager.

1. The model learned the noise of the training data.
2. It needs simpler design, more varied data and validation on new data.

How to score full marks

• Define hidden assets and give the four methods of concealment.
• Give the types and link them to the scenario.
• Group tracing methods as research, follow-up, investigation, forensic analysis, machine learning and legal tools.
• End with limits and the need to act quickly.

Common mistakes in the exam

• Listing tools with no link to the facts.
• Forgetting nominees and shell companies.
• Ignoring public records and third parties.
• Treating machine learning results as proof.

Quick revision

• Hidden: deliberately concealed to avoid legal process.
• False reporting, nominees, shells, offshore.
• Research, tracing, investigation, forensic analysis, machine learning, legal tools.
• Limits: data quality, skills, time, complexity, overfitting.

Practice questions

Questions 1 to 10 are multiple choice. Questions 11 to 13 are written. Try each before opening the answer.

Model answers

Spotted a mistake? Tell us and we will check it.