Be smart in the underwriting processes
The South African life insurance industry has skilled underwriting talent and product developers who offer products that are basically irresistible to end-users. However, the more new products become available in the market, the more pressure there is in the underwriting process.
As fraud is more prevalent today, the underwriting process risk selection is forced to be more accurate and quick, less intrusive and more cost effective to secure market leadership. Most insurers make use of traditional analytic predictive modelling data points to determine more accurate results and predict outcomes.
Fraudulent behaviour occurs at new business stage when the fraudster aims to gain maximum financial benefit within a relative short period. However, to determine the possibility of fraud the focus must turn to consumer behaviour. While the industry’s custodians of data typically have a large quantity of data, it is often too narrow in scope. Consequently models are often unable to produce the desired results.
Fraudulent activity in life cycle
Fraudsters are sophisticated in their methods and are well aware that insurers are under pressure not only to provide the best products that secure financial benefits but also race to pay policy proceeds in relative short time frames. Sadly only fraudsters and opportunist gain tremendously and it gives them their right of existence.
During the life cycle of a policy there are three direct contact points were potential fraud can be identified. These are:
- at new business stage
- during the underwriting stage
- at risk detection stage
At new business stage, traditional predictive modelling supports and assists the risk assessment. As the needs of the insured change, he or she might add or cancel the benefits of his or her existing portfolio. The focus then turns to fraudulent behaviour during the underwriting process where dynamic predictive modelling and red flag indicators are used to segment claims for assessment and validation.
Undoubtedly, the appetite for risk detection lies in the third and last contact point when a claim is submitted.
Fraud and misrepresentation will be detected and confirmed at this stage of the process with great success.
Deterring the modus operandi
Continuous data analysis is extremely important to keep in touch with the ever changing modus operandi of fraudsters. The challenge, however, remains with the portion of claims which are segmented as possible fraudulent claims, but after assessment and validation it is confirmed that the claims should be honoured. This aspect of risk detection is directly linked to reputational risk and should not be underestimated.
In the ever-important drive for the expansion of market share, insurers and underwriters need to direct and manage data collation effectively at sales stage. Insurers and assessing experts can address this data challenge by bringing additional claims data to the table.
This additional data may not only include personal data, but also demographic and other less obvious data which might require third party data partners in order to develop effective prediction and segmentation models.
Success over failure
If the data quality and quantity is acceptable, current and is being provided in a way that refreshes the model, it can result in success over failure in the prevention and detection of fraud.
For a very long time, insurers have been unwilling to spend more on acquisition costs by implementing strategies to detect fraud at the underwriting stage. The costs of assessing this, have however, become very expensive and the industry can no longer rely on the notion to identify fraudulent transactions only at the claims stage.
With increasing claims costs, insurers should improve their efforts to find new and innovative ways to reduce costs on all fronts. Rule based decision making, triangulation of data, and perhaps even the utilisation of reliable screening instruments at the sales stage, is what will determine the success or failure of insurers in future.