Sniffing Out Deception: Auto Insurance's Investigative Art

Lying to an auto insurance company is a risky business and can have serious consequences. People often lie to save money on insurance, to get more money from a claim, or to get a quote they are not eligible for. However, insurance companies have ways of finding out the truth. National databases can reveal tickets and accidents, and insurance companies can also use odometer readings to check mileage. In addition, insurance companies use outfits like Quality Planning, which use sophisticated analytical testing to uncover rating errors. The consequences of lying can include having your claim rejected, your policy terminated, facing criminal charges, and paying higher premiums.

National databases

The "Index System" or "Index Database" is a claims information service used by most insurance companies to access the claims history of every claimant. This database includes information on prior bodily injury, auto, workers' compensation, and homeowner claims, dating back 15 to 20 years. It provides details such as the coverage for which a claim was filed, the date of the incident, the type of injury, and the settlement amount. Additionally, it sends alerts to insurance companies with open claims for a particular individual.

Insurers also use two other databases, the Comprehensive Loss Underwriting Exchange (CLUE) and the Automated Property Loss Underwriting System (A-PLUS), to track and share information about their policyholders. These databases are run by outside agencies—LexisNexis for CLUE and Verisk Analytics for A-PLUS. They contain data on claims filed, including dates, causes of damage, claim check amounts, and more. This information helps insurers assess the risk associated with insuring individuals and set appropriate rates.

In addition to claims history, national databases also contain other relevant details. For example, the National Insurance Crime Bureau (NICB) created a database after the 2005 hurricane season that stores vehicle identification numbers (VINs) and boat hull identification numbers (HINs) from flooded vehicles and boats. This information is accessible to law enforcement, state fraud bureaus, insurers, and state motor vehicle departments, helping to combat salvage fraud.

Furthermore, the National Motor Vehicle Title Information System (NMVTIS) is a database that includes information on 99% of the US vehicle population. Junk and salvage yard operators, as well as insurance companies, are required to file monthly reports on vehicles declared total losses. This system helps prevent "title washing," where unscrupulous sellers switch or clone vehicle identification numbers to conceal flood damage.

Overall, national databases provide auto insurance companies with a wealth of information to verify the accuracy of applications, detect fraud, and make informed decisions about rates and coverage.

Mileage checks

There are several methods that insurance companies employ to obtain mileage data:

• Self-reporting: The insurance company may contact the insured via phone and request a current odometer reading. However, this method relies solely on the honesty of the insured and may not be verifiable by the insurer.
• Technology: Insurers may ask drivers to install tracking devices in their vehicles that automatically record and transmit mileage data to an insurance app on the driver's smartphone. Alternatively, a device plugged into the car can record mileage information directly.
• Databases: Insurance companies can cross-reference state or national databases to confirm vehicle mileage. This approach may not always accurately reflect an individual's mileage, as data is often gathered from various sources, including state inspection providers, maintenance shops, and car dealers, who may enter information incorrectly.

It is important to note that providing false information about mileage is considered fraud and can lead to serious consequences, such as policy cancellation, claim denial, increased premiums, difficulty obtaining future insurance, fines, and even criminal penalties. Therefore, it is essential to be truthful and accurate when reporting mileage to auto insurance companies.

Neighbours reporting

Insurance fraud is a serious issue, and auto insurance companies have a range of methods to detect it. One of the ways they can find out about potential fraud is through reports from neighbours or other individuals. This is especially true if you live in one state and your vehicle is registered in another. Neighbours may notice and report discrepancies, such as frequently seeing your car parked in a different area than the one you've listed as your address.

Insurance companies take address verification seriously and use various methods to confirm an individual's residential address. They may check your driver's license, vehicle registration, utility bills, lease or mortgage documents, or conduct database searches or third-party address verification services. One such third-party company is GarageConfirm, which helps insurance companies combat garaging fraud by using towers and cameras in different areas to spot your vehicle and compare how often it is seen in your listed ZIP code versus other areas.

If you provide false information about your address, there can be serious consequences. Insurance companies may deny your claim or cancel your coverage. They could also increase your premiums or label you as a high-risk driver, making it more challenging and expensive to obtain insurance in the future. In some cases, providing false information may even result in legal consequences, including fines or jail time.

Therefore, it is essential to be honest when providing information to your auto insurance company. If you have any changes to your address or other details, it is important to update your insurance provider promptly to avoid potential issues in the future.

Police reports

Insurance companies use police reports to investigate and resolve insurance claims. They can determine how much each party is at fault and adjust premiums accordingly. For example, if one driver is found to be 70% at fault, their insurance company will pay 70% of the bill, while the other driver's insurance will cover the remaining 30%. In some cases, insurance policies may even pay the full amount to the driver who is less at fault.

In addition to police reports, insurance companies also have access to national databases that track driver information, such as traffic tickets and accidents. This information can be used to verify the details provided by the insured and identify any discrepancies or lies. Insurance fraud is considered a crime, and those who are caught lying on their insurance applications or claims may face serious consequences, including higher premiums, loss of insurance coverage, or even legal penalties.

Overall, police reports play a crucial role in the auto insurance claims process, helping insurance companies make fair decisions and protect themselves from fraudulent claims.

Analytical testing

Text Analytics and Data Mining

AI and machine learning algorithms play a pivotal role in text analytics and data mining, enabling the processing of vast amounts of structured and unstructured data from various sources. This includes mining insurance applications, claim forms, and adjuster notes for red flags and inconsistencies. Natural Language Processing (NLP) is a key tool in this process, aiding in the identification of fraudulent claims by analysing adjuster remarks and historical claims data.

Logistic Regression

Logistic regression is a statistical tool used to model the relationship between a dependent variable and a set of independent variables. In the context of insurance fraud detection, the dependent variable is typically binary, indicating whether a claim is fraudulent or legitimate. Logistic regression helps identify patterns and variables associated with fraudulent claims, such as claims history, policy details, and vehicle information.

Predictive Analytics

Predictive analytics is a powerful tool for proactive fraud detection, utilising advanced data mining and behavioural pattern analysis. By analysing policyholder profiles, behaviour patterns, and touchpoints, insurance companies can assess the fraud risk associated with each claim. This allows for early warning systems and more informed decision-making, reducing financial losses and improving customer satisfaction.

Machine Learning Techniques

Machine learning algorithms are trained on extensive datasets to identify patterns and anomalies in insurance claims. These algorithms become more accurate as they process more data, enabling the detection of outliers and suspicious requests in real time. Common machine learning techniques include decision trees, support vector machines, Bayesian networks, and neural networks.

Graph-based Approaches

Graph-based fraud detection systems, such as InfDetect, utilise graph theory and network analysis to uncover relationships and connections between potential fraudsters. By modelling the interactions between buyers, sellers, and other entities in a large-scale graph, insurance companies can identify organised fraud rings and suspicious activities, leading to more effective fraud detection and prevention.

Model Comparison and Validation

Comparing the performance of different analytical models is essential for selecting the most effective approach. This involves evaluating measures like Pearson Chi-Square, log-likelihood, and pseudo R-square values to determine the best-fit model for fraud detection. Additionally, hold-out samples or cross-validation techniques can be employed to test the model's predictive accuracy and generalisability.

In conclusion, analytical testing in auto insurance fraud detection encompasses a range of advanced techniques, from text analytics and logistic regression to machine learning and graph-based systems. By leveraging these tools, insurance companies can identify deceptive behaviours, protect honest policyholders, and maintain the integrity of the insurance industry.

Frequently asked questions