Head of Science, Global Casualty
AIG plots 100 percent low-risk growth, avoids $75 million in losses
With pressure on operating expenses, escalating regulatory pressures, increasing competition and the cost of paying out claims, even insurance companies need a little, well, insurance. For AIG, a leading property and casualty insurance organization serving more than 45 million clients in over 160 countries, some of that insurance comes from using SAS® Business Analytics.
According to Head of Science, Global Casualty John Savage and Head of Science, Global Property David Lee, SAS has become the analytics backbone of the Strategic Risk Analysis Group at AIG. Their group takes on projects in an effort to continually optimize business performance. Their use of 16 underwriting and finance predictive models has helped prevent millions of dollars of potential future losses on an insurance and reinsurance portfolio of approximately $13 billion.
... the tools give us a real-time review of our portfolio. We are now much more confident in making reinsurance decisions.
Three areas of the insurer's business in particular highlight the payoff that the Strategic Risk Analysis Group created: executive liability insurance, catastrophe planning, and financial accounting.
Assessing executive liability risk
To help underwriters assess insurance risk in the area of executive liability insurance, Savage and Lee's team statistically analyzed hundreds of potential loss drivers by researching the filings of every public company since 1996. Identifying six significant predictors, the team deployed its own quantitative risk model (QRM), a web-based, on-demand tool that summarizes risk profiles and enables risk-based business decisions. In an 18-month period, AIG used the solution to target $14 million in new, executive liability business, representing 100 percent growth in that segment. In addition, the modeling tool helped prevent a potential loss of $75 million from certain executive liability accounts over the course of a year.
Better planning for catastrophes
No one knows when catastrophes will strike, and they come in many different forms. But due to AIG's use of analytics, the company now has a clearer picture of its property risk exposures in many geographies, which provides a higher level of confidence in the insurance risk decisions it makes.
"We had relied on reinsurance as a backstop to some degree because we thought we might be overexposed, but we weren't quite sure," explains Savage. "We wanted to minimize that, so the tools give us a real-time view of our portfolio. We are now much more confident in making reinsurance decisions. Today we have a daily view of our risk profile. What we didn't know before was how densely our policies were concentrated in certain locations, like California for earthquakes or Miami for hurricanes. If there was an event in either of those areas, we didn't immediately know what our exposure was; now we do. So, when we go to purchase reinsurance, we know more precisely where our risks are located and how much reinsurance to buy."
Strategic financial planning
When a calamity does strike, the insurance company is under pressure to quickly respond to the needs of affected clients. To ensure that funds are accurately accounted for in the company general ledger for auditing purposes, the analytics team built an automated reconciliation tool using SAS. Due to the insight the tool provides, the company is able to reduce the amount of funds it must keep in reserve to cover discrepancies related to unreconciled payments. In addition, Savage says the tool allowed the company and its auditors to look back over a number of years and, using a data matching process, find millions of dollars in unreconciled payments, providing the company with a deferred tax credit of $10 million.
To help with finding the tax credit, Savage says the company’s auditors used the SAS Add-In for Microsoft Office to access the raw data – for hundreds of millions of transactions.
"They couldn't easily look at this kind of data, but once we put it in a SAS data set in the SAS® Scalable Performance Data Server® and then put it through the add-in for Microsoft Excel, we could leverage the power of SAS with outside consultants," he explains. "They were finding transactions left and right. It was a huge win for us. Many non-SAS users love the tool because they can look at 100 million records in Excel; they're blown away by that."
"SAS was critical to the reconciliation," adds Lee. "With hundreds of millions of records, previous attempts to do the reconciliation using a variety of software were too slow because our accountants couldn't see the whole data set. SAS has virtually no limit in the number of rows that you can look at."
In another project, the Strategic Risk Analysis group used SAS to build probability-based exposure models to estimate bad-debt reserve needed for uncollected premium receivables, based on open balances across multiple lines of business. The methodology and algorithms comply with audit requirements and provide stable exposure estimates each quarter.
With the Strategic Risk Analysis team well-established and adding considerable business value at AIG, the group is regularly approached to take on new projects. According to Savage, the projects must be qualified and contribute to increasing profitability.
"Modeling with SAS is highly scalable. If we get to a point where something looks like an opportunity, we run it by executive management and show them the value. The larger the organization, and the more data there is, the more opportunity there is to add value with SAS," said Savage.
AIG required an analytics platform to estimate the risk of future loss, help underwriters assess and price insurance risk, and reconcile claims payments, as well as estimate bad-debt reserve funds for premium receivables.
- $14 million in new, low-risk business, representing 100% segment growth.
- Avoided a potential loss of $75 million from certain executive liability accounts.
- $10 million deferred tax credit.
- Reduced requirement for bad-debt reserve funds.
- Accurate, current account reconciliation process.
- Real-time view of risk exposure.
- Ability to analyze hundreds of millions of records.
- Sixteen underwriting and finance predictive models, supporting $3.2 billion in underwriting premiums and reinsurance annually.