Ratemaking: It only takes one person
Ohio Mutual powers its ratemaking with SAS® BI for Midsize Business
Ohio Mutual writes nearly $190 million in annual premiums through independent agents in seven states, serving homeowners, drivers and farmers. Using SAS® Enterprise Guide®, Ohio Mutual can quickly forecast rate changes all the way down to the policyholder – with just one dedicated analyst.
That analyst is Bob Roesch. He is quick to stress that he is not an IT specialist by background, nor a statistician, but rather an actuary. “An insurance man through and through!” he jokes. Nonetheless he has considerable confidence in SAS: Roesch advised the company to invest in the software, and Ohio Mutual has been using it since shortly after he joined the mid-sized insurer in 2009.
If you are a midsize insurance company like we are, one person can handle data processing tasks quite efficiently in SAS.
Ohio Mutual runs SAS on its mainframe, where it has a DB2 database that holds roughly 10 million records in its main insurance files. “The tables are usually quite detailed, and about 85 percent of my time is spent compiling information so that we can do the ratemaking. I use SAS to clean up the data – it’s a good language for data manipulation tasks – and then we are ready to do the analysis, mostly using SAS BI for Midsize Business with SAS Enterprise Guide,” Roesch says.
Ratemaking is a core process in the insurance business. “We always have rates to do. In principle, it’s simple. We earn premiums and we process claims. So we have to compare the two to determine rates. In practice, it’s complex and time consuming because to maximize profitability, every time you need to change the rate, you have to recalculate all of the premiums,” he says.
“In the old days, when that was done manually, you needed to employ a lot of people to calculate the rates. Even using Excel, it is a horrendous workload if you have to calculate say 100,000 new policy premiums. But with SAS, once the programs are set up, it all runs very fast and effectively, and we can calculate at a high level of detail.”
An important part of the equation is calculating losses. At Ohio Mutual, these arrive in detailed flat files and have to be compiled and paired with the right policies. “There's another thing that actuaries have to do that some companies find difficult,” Roesch says. “Losses have to be put into a triangular form because in property and casualty insurance, many claims are not settled until well after the policies have expired. Usually, there is a trend or a pattern in the elapsed time, and we have to account for that by constructing development triangles. Again, by using SAS, this tricky problem becomes quite easy because it’s essentially the same procedure that has to be run over and over again.
“If you have a midsize insurance company like we are, one person can handle it quite efficiently in SAS,” Roesch adds. Before he arrived at Ohio Mutual, the company used an expensive software the sole purpose of which was to produce loss triangles. “The software had nowhere near the flexibility that I have using SAS.”
Once Roesch has analyzed the losses and calculated rates, the results go over to the sales team. “We show them what's happened to these groups of policies, and they can make decisions on pricing from that. Sure, they often apply some sixth sense on what rates the customer is prepared to accept, but they are making decisions based on facts.”
The final decision can sometimes take a couple of iterations, and Roesch will need to do some forecasting with SAS BI for Midsize Business. He mainly uses the GENMOD generalized linear modeling procedure to project one to two years forward, which is the rate effective period. “We don’t always have that much data to go on, but we get very effective results,” he says.
Since Roesch is the only dedicated SAS analyst, he needs to output the results of his analysis to Excel, where the final rates can be shared with underwriters and project managers. “This is really easy with SAS Enterprise Guide,” he says.
“In general, SAS Enterprise Guide is menu-driven so you don’t have to remember code, which means the whole process is manageable from end to end. We are bringing a couple of new users in house, and they are getting up to speed pretty quickly.”
In summary, Roesch cites three key benefits of using SAS at Ohio Mutual.
First of all, data access: “Without it we would need to employ someone to pull the data using SQL software and put it into a spreadsheet. You would also need to summarize the data and make approximations, which means you couldn’t get the same level of detail. That’s how it was done in the old days, but with SAS you can extract and use data at a greater level of granularity. Your ratemaking is therefore more accurate by a factor of a few percentage points, which can be the difference between profit and loss.
“The second major benefit is speed of analysis and information delivery. I know because I have seen what happens at other companies of similar size using a staff of five analysts – just because they didn't have the right software.
“The third benefit is the quality of the analysis. We can build in all kinds of factors for different classes of policy holders, geography and so forth. For me, this is vital because without quality information, you're always going to struggle.”
Time-consuming manual processes, limited staff and siloed data.
More efficiency in ratemaking tasks, improved access to data sources, greater accuracy in calculating losses and predicting rates, and faster information delivery and decision making.