We have seen how Predictive Modeling and Price Optimization has worked in the past few years. Company after company losing their shirt on Auto after going this route. At what point do they look at their book and say, what is going on here? The goofy actuary looks at the board and tells them, it should work because my numbers said it would. Back to the drawing board. By the way, re-scoring, re-tiering and state wide rate increases are not the way to retain business.
Maybe I missed something. Which companies,of any significance, are “losing their shirt”? Seems like the major personal auto carriers, with the exception of State Farm (who rarely writes under a 100 CR on personal auto) has reported profitable results.
SW, does the name GEICO resonate with you? Just reported a $4.4 billion underwriting loss on Personal Auto. Many others like Travelers, Safeco are also reporting losses on Auto. That is why they are taking more rate in the near future. State Farm is a prime example of losing on Auto.
Agent – I have Warren Buffett’s annual letter to shareholders in front of me and he reports a $460 million underwriting profit for GEICO in 2015. Yes, that is down from previous years, but still profitable. And that is not their operating profit, just their underwriting profit. With over $15B in “float” they have a very health underwriting profit. Where did your data come from? Also, State Farm typically has an operating profit on Auto despite the underwriting loss. As a mutual, they seem to be happy with their result.
Agent was asked for the source of his data – why am I not surprised we all hear the sound of crickets now?
March 7, 2016 at 8:11 pm
UW says:
Like or Dislike:
3
1
His source is his genius brain, all other data is invalid in his mind. He’s never revealed a source. He will vaguely say it was on a site, and then attack you for not finding it.
He is oddly focused on the auto industry too.
March 2, 2016 at 2:42 pm
Jadefox says:
Like or Dislike:
8
0
I’ve been in rate review meetings during which the actuary proclaims that auto needs 46% increase in rate to get the book to profitability.
Then the actuary sits back and watches as underwriting and marketing VP’s begin the dance.
Well, we can’t get that much. Our agents will abandon us for other markets. Can be done, says marketing.
Underwriting says, well, we need something, so let’s go with 6.5%. We should be able to keep our business then. We’ll just underwrite new business more closely.
The actuary then says, you’re compounding the problem. The rate need last year was 32% and you took 3%. Same arguments were made and we are still losing money.
In unison, the VP’s respond, we will do a better job getting our appetite out to the agents so they will only send us what we want to write.
And so it goes, until the CEO calls up and says, “DUMP THE AUTO BOOK”.
So what you are saying is that it is not that there is a problem with predictive modeling, just the lack of execution when incorporating what the model states.
Sounds like it is the decision makers that have resulted in poor auto ratios, not predictive modeling.
Yes Sir! Imagine what would happen if Carrier X walks into the Agent/Brokers office to let them know to expect substantial increase in Auto rates. Then the weeping, wailing and gnashing of teeth begins. A/B threatens to “move the entire book”! And so it goes.
Why, because the high up CEO and VP’s declare a growth year and impose substantial growth objectives, while at the same time, imposing the need for profit. Never ending conflict.
The A/B’s don’t care as long as they retain the account. Predictive modeling will not end the problem when it always comes down to price.
Remember, one company is not that much different from the other and any proclaimed differentiation is used to support the higher price.
Once I was in a meeting when a senior VP declared, yes, we are higher so remind broker that placing the business with us will result in higher commission!!!! As if the broker could not figure that out for himself. Did it work? NEVER.
Jadefox, I have been in this business for a long time and the Pre-Predictive modeling years and Pre-Price Optimization years were very good and for the most part, rates were fairly stable. We could sell service, great coverage at a fair price. We now have descended into a crap shoot with blanket rate increases, re-scoring, re-tiering, maintenance rates, discrimination against certain classes of insureds etc. Did you see the article where Renters were surcharged more by most carriers with Liberty Mutual being the biggest offender? It is very hard for agents to deal with the idiot Predictive Modelers and Actuaries who have never sat down with a client and tried to explain why their rates went up 20-25% when they had good credit, paid their bills on time and didn’t have any claims. We often have to move it or lose it and hope the one we move it to doesn’t do the same thing at next renewal.
March 3, 2016 at 3:04 pm
Yogi Polar Berra says:
Like or Dislike:
4
0
The ‘debate over needed rate and market rate’ reminds me of the following cliche/ conundrum:
If all your friendly market competition jumped off a bridge, would you?
42 years in this silly business. Been through the ups and downs. The same “senior executives” that start clamoring for GROWTH are the same ones that call a HALT with things get bad.
Stock companies are the worst. Got to maintain the stock value so the executives can exercise their options.
I’m now retired and glad of it.
No wonder young folks are not attracted to underwriting. They work so hard only to be beat to death in never endiing underwriting audit, dealing with conflicting messages and then a performance review for not making goal, even if the company told them to dump millions.
Jadefox, the same thing goes on with Commercial as well as Personal Lines. I recently had a renewal on a 10 unit fleet come up for renewal and it came in at 25% over expiring for a loss free account. I called up the underwriter and told him he was running the account off and asked for relief. Answer: Sorry, can’t help you. Ask the actuary why we went up. A week later, the marketing rep called and said they wanted to write business and to give them a try. I related my story and she had no reply. Egg was dripping all over her face.
March 2, 2016 at 11:22 am
Concerned Modeler says:
Like or Dislike:
12
7
Seems like a lot of the comments on this online journal are usually from agents and typically non-captive ones.
I’m not sure how people from organizations like Kemper for instance who simply broker other companies’ products would know anything at all about the “nuts-and-bolts” of the ratemaking process.
Aside from Marketing models aimed at assisting these agents in targeting different demographic groups for their agents’ book of business I don’t anticipate these 3rd party insurance groups having much analytical rigor in their operations.
Concerned, you don’t seem too concerned. Perhaps you are too removed from what is going on in the marketplace. I am amazed at how so many companies managed to succeed and grow prior to Predictive Modeling and Price Optimization coming on the scene. I wonder why Price Optimization schemes have been under attack in a number of states and are due to be banned. These schemes were all thought up by dufus modelers and actuaries. They screw the public big time.
Since you are such an expert, how many companies have had their Price Optimization schemes approved by states? It sounds like they are all getting, rightfully, disapproved by regulators.
In addition, for someone who is so anti-regulation and pro free market, you should be for whatever the companies decide on how they rate and/or underwrite.
This is not a hard concept to grasp – the entire principle of insurance is based on predictive modeling!! The insurance industry is considered a part of the financial sector! Actuarial studies occur ALL the time! Nothing has changed about insurance and it’s principles, except that new words have been assigned to functions.
Agent, I couldn’t agree with you more. How did the top three build the largest insurance delivery systems in the free world before consultants took over the underwriting departments. So much about the insurance process has been relegated to persons who never put a policy in a clients hand. The math is the same as it was 15 years ago, we are just wasting more premium paying non performers for their opinions.
I don’t know that all is the same as 15yrs ago. The legal environment is much more aggressive and “new” injuries are being created everyday. Combine that with increased costs and it’s hard to stay ahead of the pricing curve.
What does putting a policy in the hand of a client have to do with underwriting risks as targeted groups that are more profitable than other groups?
Predictive Models and Consultants that develop them are much more qualified to identify the criteria that will yield the best overall results than field underwriters who are pre-occupied with sales.
Increased use of Predictive Models, and reduced use of Price Optimization Models / Processes, will result in better matches of price and risk than the current system where too much underwriting judgement used to chase premium volume results in compressed pricing schemes. Socialist politicians will complain, but high rated risks will present challenges to reduce underlying costs that have been faced before and successfully resolved. It’s better to reduce high costs than subsidize them through ‘underwriting judgement’ reduction. That is likely to eventually lead to greater market stability and price affordability for all risks.
You are right on. Most of these have very little analytic rigor, and there isn’t a lot of need for it.
You should be aware agent doesn’t believe in statistics. He thinks the Farmer’s Almanac is more accurate than weather forecasters, the earth isn’t warming, BLS figures are wrong, and on and on. He is also incompetent when it comes to statistics. For example, he once complained saying something along the lines of weather forecasters are dumb, because their stupid models call for a 40% chance of rain, and then it barely rains, or rains all day, not 4/10 of the day.
I see some companies apply it successfully. Liberty, for instance, in the northwestern states appears to be very successful. For smaller companies the challenge is to have enough data points for credible results. The downside for these markets is that the complicated models are not available from large competitors in many states. Before predictive modeling, small markets could obtain rates from larger entities from their regulatory filings.
Liberty advertises a lot and the Brad girl is cute, but did you know that Liberty has the highest charges for Auto for Renters among all the major companies? It was on Property Casualty 360 recently. It seems they penalize the lower credit score of renters much more than State Farm, Allstate, Farmer’s group. Now that is creative modeling, isn’t it?
I work for a mid-size regional company. Predictive modeling has helped us become profitable in Personal Lines. It’s allowed us to expand our appetite for what was borderline business and reduced the adverse selection we would see from the larger companies. Also, recent profitability issues some companies are experiencing are not due to predictive modeling, but to cars being on the road more because gas in cheap. After years of seeing average auto losses go down, they moved up. Rates will follow.
Please explain how writing borderline business is a good thing. Kemper thought they could expand their appetite for borderline non standard Auto and it ate their lunch.
It is really quite simple. There is a saying among intelligent insurance professionals, “There is no such thing as a bad risk, only a bad rate”. If predictive modeling helped that company find a good rate for that borderline risk, then they can expand their appetite and still be profitable.
Ron is right. Plus, with new data and new techniques for analysis they can see, for example, a class that may have been totally excluded or largely excluded because it was borderline, but actually can be written because they have new ways of estimating the probability of a loss based on data they couldn’t use before.
Also, if they can use data to safely write borderline business little better they have a whole new pool of money to invest. Even if they lose a little it’s like using an interest free loan to invest with and earn profit. Long-term the yearly return on the S&P 500 is 7% in an ETF.Even if you end up losing 2% on a new borderline venture it can be worth it.
I’m still waiting for the first comment from Agent showing anything beyond the most superficial understanding of the insurance industry.
knowall, I believe the experienced underwriters who actually thought, underwrote on time tested principles were more accurate than the computer models used by goofy actuaries and modelers.
Agent – question for you. (and this isn’t an attack, I really am curious as to what an agent thinks.) 30 years ago when I was a PL Underwriter with a “preferred” company, I was told to never write someone who was divorced in the past year. Now, many companies have abandoned those types of rules in favor of developing a rate(using predictive analytics and additional rating variables)and charging the right rate for such a risk. As an agent, would you rather have the conversation with a recent divorcee about why you cannot write them or tell them you can write them at X price?
CL, thank you for not attacking me and wanting to know how an agent thinks. I have been in this business for over 30 years and have seen every soft and hard market and all the swings up and down. Personally, I have not had any of my preferred markets ever tell me I couldn’t write a divorcee. If it was a current customer, we often had to split up the coverage and write both separate policies. Sometimes, the credit score suffers and a new applicant pays slightly higher premiums. The big problem is when an irate husband or wife tries to take the others car off the policy before the divorce is final. That can be dicey. I would bet a Predictive Modeler couldn’t answer that since they have never dealt with the public before.
Really Milner? Our Preferred markets have always done better than the Standard or uprated market. We throw problem accounts to Progressive since they will write about anyone for a price. No need to ruin the good market with substandard business who don’t pay bills on time and have more claims.
Predictive modeling is really a combination of underwriting and ratemaking on steroids. Over the years, many fine underwriters and actuaries noticed trends on general classification like divorce people or young people in sports cars and made rules and rate classes. A sound predictive model can refine those classifications in more detail and constantly update them. A trend that might have been valid 40 years ago may not be valid today.
Have some experience working with ‘friend of the court’ regarding child support calculations — they like to pull out a chart, while sitting in an air conditioned desk environment, and say,” okay your gross income is “x,” and you have 3 kids, therefore you have to pay “y.” Challenged them once about the fairness and rationality of this method and the clerk responded, “take your comments to your lawyer,” and hung up on me.
I guess using these predictive models without any (ongoing) field or underwriting expertise/input could be a mistake.
You are correct knowall. Using the models without any judgment is typically not the recipe for success. Sometimes the models generate indicated paths that cannot be followed due to social, regulatory, competitive and good ole common sense reasons. The best companies know when it is necessary to deviate from the model.
Cl, sometimes the best companies make big mistakes in judgment on their modeling. One of ours did that in 2014 and implemented state wide rate increases, re-tiered their book, re-scored all the customers of their book. Guess who took the heat from all those changes? Net result was that many customers told us to move them or they would find someone else.
It is an interesting question if field or underwriting expertise would optimize the predictive modeling process. I have been in the business for 30 years and wonder if some of my past perceptions may hold me back in this process. In the past we would have special underwriting rules and judgments to compensate for the lack of flexibility in our rating plans and underwriting guidelines. The predictive modeling process, in theory with valid and credible data, should eliminate the bias and indicate the appropriate rate for a risk.
Observer, the key word is “valid”. Often, the modeler data is both flawed and biased. How valid is discriminating against older drivers who are loss free, pay their bills on time and giving them a rate increase for no apparent reason? How valid is Price Optimization schemes designed to raise rates on long term loyal loss free customers. See Allstate for that one and it is happening with other markets as well.
Older drivers have ever-decreasing vision and ever-slowing reaction times. They are more likely to be involved in accidents at 90 than they were at 50.
The problem with your statement is that it usually starts about age 65 with many markets on discrimination by age. Most people are dead by age 90 or haven’t been driving for many years and in a nursing home. Customers we have at that age(65-75) are sharp, have glasses for good vision, don’t have accidents, pay their bills on time and all around good customers who give no one any trouble.
And if you tested those 65-75 year old drivers on how perceptive they are and how quick they are able to react, you would find they don’t have the same clarity of vision or are as fast as reacting than they did at, say, age 35.
This is why older drivers are charged more — they do not see as well as they used to, their reaction times are slower, and those are only going to get worse as they continue to age.
March 16, 2016 at 12:15 am
UW says:
Like or Dislike:
1
0
This is just statistically incorrect. “Many,” “some,” “my clients,” etc are irrelevant; the only thing that matters is the population on a whole which makes up the risk pool. This age group is without debate a group that is more dangerous than others and gets increasingly worse every day.
March 5, 2016 at 8:28 am
SWFL Agent says:
Like or Dislike:
4
1
“Mature” drivers can be profitable. However unlike young drivers whose risks characteristics typically improve over time, a senior’s driving skills will diminish over time. Too often, we have senior drivers that will not stop driving until repeated incidents occur and it’s difficult for carriers, from a pricing perspective, to know when this time occurs.
You know what Rosenblatt and SWFL, not too many years ago, the market thought mature drivers were a favored class to write. They typically hardly ever have a claim, pay their bills and don’t shop their coverage much. I would much rather have one of these accounts than the 25-35 age class who is so distracted by their cell, they run over someone.
March 7, 2016 at 11:56 am
Ron says:
Like or Dislike:
0
2
Agent,
In other words, you want the easiest accounts you can get. I wonder if your carriers know this is truly how you think. It may be high time for them to terminate their contract with you and move onto an agent who does not mind putting in the work.
March 7, 2016 at 12:25 pm
Rosenblatt says:
Like or Dislike:
3
1
Agent – didn’t you have a personal experience where you had to take your father’s (or grandfather’s, I don’t fully recall) license away? Why did you do that? Did they keep crashing their car or they couldn’t safely drive at night because they couldn’t see so well in the dark anymore? Those are just two reasons why aging drivers see increases in premium even though they didn’t experience a loss — as people get older, they will become increasingly less safe behind the wheel. Do you disagree with me on that?
March 7, 2016 at 5:57 pm
Agent says:
Like or Dislike:
1
1
Ron, the cubicle knat. Had you ever been a real agent, you may have a different attitude about insured’s and the market. I have been putting in the work for a long time, you idiot. By the way, why would the carriers I do business with terminate my contract when I produce consistently profitable business for them and they reward me with excellent contingency each year? We set a new record for profitability for 2015. That is something you will never see. Are you jealous of successful agents or what is it?
Does this mean that you will actually pay your taxes this year or do I have to pay them for you like the past several years.
We do well in spite of the Obama economy, not because he has done anything to improve it. We work twice as hard to support the leaches of society. Our carriers do like us, however.
Damn, nice one Ron; although it’s not hard, it’s funny seeing this clown made to look foolish.
He says he prefers the accounts that don’t switch, because the younger people cause more accidents (texting of course), you basically quote him and he denies it. Total clown.
Of course, he’s wrong on the facts too. Younger drivers in the age he cites are slightly more dangerous if you don’t control for urban/rural. But the age he specifically doesn’t want is the exact age they start to become safer and the age he prefers is when they begin to get increasingly more dangerous every year-unsurprisingly. I say unsurprisingly because he had been totally wrong on literally every topic I’ve seen him write about. He cannot agree with data and modeling because he’s so divorced from reality it would challenge every single thing he thinks and he couldn’ttalk handle it.
As for the data modeling, he would be bankrupt almost immediately if he actually had to figure out how much to charge to cover losses, what business to write, etc. Just 100% clueless, and qualified for no other role than that as a salesman where blabbing matters more than actual reality, I don’t believe for a second he could last a month at a legit agency.
Rosenblatt, No, I didn’t have to take either my fathers or mothers car away from them since they died. Some seniors do have a problem at some point seeing. Have you ever heard of a Medical Statement being filled out to list conditions? It is pretty common and starting about age 75, some companies require them. If there are no compelling reasons to list health concerns, it is generally ok to let them drive. By the way, they are far safer than the young cell crowd who are addicted to their phone and often text while driving. I see it every week and it is disgusting.
Hidden due to low comment rating. Click here to see.
Maybe I missed something. Which companies,of any significance, are “losing their shirt”? Seems like the major personal auto carriers, with the exception of State Farm (who rarely writes under a 100 CR on personal auto) has reported profitable results.
SW, does the name GEICO resonate with you? Just reported a $4.4 billion underwriting loss on Personal Auto. Many others like Travelers, Safeco are also reporting losses on Auto. That is why they are taking more rate in the near future. State Farm is a prime example of losing on Auto.
Agent – I have Warren Buffett’s annual letter to shareholders in front of me and he reports a $460 million underwriting profit for GEICO in 2015. Yes, that is down from previous years, but still profitable. And that is not their operating profit, just their underwriting profit. With over $15B in “float” they have a very health underwriting profit. Where did your data come from? Also, State Farm typically has an operating profit on Auto despite the underwriting loss. As a mutual, they seem to be happy with their result.
Agent was asked for the source of his data – why am I not surprised we all hear the sound of crickets now?
His source is his genius brain, all other data is invalid in his mind. He’s never revealed a source. He will vaguely say it was on a site, and then attack you for not finding it.
He is oddly focused on the auto industry too.
I’ve been in rate review meetings during which the actuary proclaims that auto needs 46% increase in rate to get the book to profitability.
Then the actuary sits back and watches as underwriting and marketing VP’s begin the dance.
Well, we can’t get that much. Our agents will abandon us for other markets. Can be done, says marketing.
Underwriting says, well, we need something, so let’s go with 6.5%. We should be able to keep our business then. We’ll just underwrite new business more closely.
The actuary then says, you’re compounding the problem. The rate need last year was 32% and you took 3%. Same arguments were made and we are still losing money.
In unison, the VP’s respond, we will do a better job getting our appetite out to the agents so they will only send us what we want to write.
And so it goes, until the CEO calls up and says, “DUMP THE AUTO BOOK”.
Jadefox,
So what you are saying is that it is not that there is a problem with predictive modeling, just the lack of execution when incorporating what the model states.
Sounds like it is the decision makers that have resulted in poor auto ratios, not predictive modeling.
Yes Sir! Imagine what would happen if Carrier X walks into the Agent/Brokers office to let them know to expect substantial increase in Auto rates. Then the weeping, wailing and gnashing of teeth begins. A/B threatens to “move the entire book”! And so it goes.
Why, because the high up CEO and VP’s declare a growth year and impose substantial growth objectives, while at the same time, imposing the need for profit. Never ending conflict.
The A/B’s don’t care as long as they retain the account. Predictive modeling will not end the problem when it always comes down to price.
Remember, one company is not that much different from the other and any proclaimed differentiation is used to support the higher price.
Once I was in a meeting when a senior VP declared, yes, we are higher so remind broker that placing the business with us will result in higher commission!!!! As if the broker could not figure that out for himself. Did it work? NEVER.
Jadefox, I have been in this business for a long time and the Pre-Predictive modeling years and Pre-Price Optimization years were very good and for the most part, rates were fairly stable. We could sell service, great coverage at a fair price. We now have descended into a crap shoot with blanket rate increases, re-scoring, re-tiering, maintenance rates, discrimination against certain classes of insureds etc. Did you see the article where Renters were surcharged more by most carriers with Liberty Mutual being the biggest offender? It is very hard for agents to deal with the idiot Predictive Modelers and Actuaries who have never sat down with a client and tried to explain why their rates went up 20-25% when they had good credit, paid their bills on time and didn’t have any claims. We often have to move it or lose it and hope the one we move it to doesn’t do the same thing at next renewal.
The ‘debate over needed rate and market rate’ reminds me of the following cliche/ conundrum:
If all your friendly market competition jumped off a bridge, would you?
42 years in this silly business. Been through the ups and downs. The same “senior executives” that start clamoring for GROWTH are the same ones that call a HALT with things get bad.
Stock companies are the worst. Got to maintain the stock value so the executives can exercise their options.
I’m now retired and glad of it.
No wonder young folks are not attracted to underwriting. They work so hard only to be beat to death in never endiing underwriting audit, dealing with conflicting messages and then a performance review for not making goal, even if the company told them to dump millions.
Jadefox, the same thing goes on with Commercial as well as Personal Lines. I recently had a renewal on a 10 unit fleet come up for renewal and it came in at 25% over expiring for a loss free account. I called up the underwriter and told him he was running the account off and asked for relief. Answer: Sorry, can’t help you. Ask the actuary why we went up. A week later, the marketing rep called and said they wanted to write business and to give them a try. I related my story and she had no reply. Egg was dripping all over her face.
Seems like a lot of the comments on this online journal are usually from agents and typically non-captive ones.
I’m not sure how people from organizations like Kemper for instance who simply broker other companies’ products would know anything at all about the “nuts-and-bolts” of the ratemaking process.
Aside from Marketing models aimed at assisting these agents in targeting different demographic groups for their agents’ book of business I don’t anticipate these 3rd party insurance groups having much analytical rigor in their operations.
Hidden due to low comment rating. Click here to see.
Agent,
Since you are such an expert, how many companies have had their Price Optimization schemes approved by states? It sounds like they are all getting, rightfully, disapproved by regulators.
In addition, for someone who is so anti-regulation and pro free market, you should be for whatever the companies decide on how they rate and/or underwrite.
This is not a hard concept to grasp – the entire principle of insurance is based on predictive modeling!! The insurance industry is considered a part of the financial sector! Actuarial studies occur ALL the time! Nothing has changed about insurance and it’s principles, except that new words have been assigned to functions.
Agent, I couldn’t agree with you more. How did the top three build the largest insurance delivery systems in the free world before consultants took over the underwriting departments. So much about the insurance process has been relegated to persons who never put a policy in a clients hand. The math is the same as it was 15 years ago, we are just wasting more premium paying non performers for their opinions.
I don’t know that all is the same as 15yrs ago. The legal environment is much more aggressive and “new” injuries are being created everyday. Combine that with increased costs and it’s hard to stay ahead of the pricing curve.
What does putting a policy in the hand of a client have to do with underwriting risks as targeted groups that are more profitable than other groups?
Predictive Models and Consultants that develop them are much more qualified to identify the criteria that will yield the best overall results than field underwriters who are pre-occupied with sales.
Increased use of Predictive Models, and reduced use of Price Optimization Models / Processes, will result in better matches of price and risk than the current system where too much underwriting judgement used to chase premium volume results in compressed pricing schemes. Socialist politicians will complain, but high rated risks will present challenges to reduce underlying costs that have been faced before and successfully resolved. It’s better to reduce high costs than subsidize them through ‘underwriting judgement’ reduction. That is likely to eventually lead to greater market stability and price affordability for all risks.
You are right on. Most of these have very little analytic rigor, and there isn’t a lot of need for it.
You should be aware agent doesn’t believe in statistics. He thinks the Farmer’s Almanac is more accurate than weather forecasters, the earth isn’t warming, BLS figures are wrong, and on and on. He is also incompetent when it comes to statistics. For example, he once complained saying something along the lines of weather forecasters are dumb, because their stupid models call for a 40% chance of rain, and then it barely rains, or rains all day, not 4/10 of the day.
I see some companies apply it successfully. Liberty, for instance, in the northwestern states appears to be very successful. For smaller companies the challenge is to have enough data points for credible results. The downside for these markets is that the complicated models are not available from large competitors in many states. Before predictive modeling, small markets could obtain rates from larger entities from their regulatory filings.
Liberty advertises a lot and the Brad girl is cute, but did you know that Liberty has the highest charges for Auto for Renters among all the major companies? It was on Property Casualty 360 recently. It seems they penalize the lower credit score of renters much more than State Farm, Allstate, Farmer’s group. Now that is creative modeling, isn’t it?
I work for a mid-size regional company. Predictive modeling has helped us become profitable in Personal Lines. It’s allowed us to expand our appetite for what was borderline business and reduced the adverse selection we would see from the larger companies. Also, recent profitability issues some companies are experiencing are not due to predictive modeling, but to cars being on the road more because gas in cheap. After years of seeing average auto losses go down, they moved up. Rates will follow.
Please explain how writing borderline business is a good thing. Kemper thought they could expand their appetite for borderline non standard Auto and it ate their lunch.
Agent,
It is really quite simple. There is a saying among intelligent insurance professionals, “There is no such thing as a bad risk, only a bad rate”. If predictive modeling helped that company find a good rate for that borderline risk, then they can expand their appetite and still be profitable.
Ron is right. Plus, with new data and new techniques for analysis they can see, for example, a class that may have been totally excluded or largely excluded because it was borderline, but actually can be written because they have new ways of estimating the probability of a loss based on data they couldn’t use before.
Also, if they can use data to safely write borderline business little better they have a whole new pool of money to invest. Even if they lose a little it’s like using an interest free loan to invest with and earn profit. Long-term the yearly return on the S&P 500 is 7% in an ETF.Even if you end up losing 2% on a new borderline venture it can be worth it.
I’m still waiting for the first comment from Agent showing anything beyond the most superficial understanding of the insurance industry.
Remember, statistical analysis is not a substitute for thinking.
IT is, however, a substitute for anecdotal “analysis”, which is often what people resort to in response.
Statistical analysis IS the thinking part.
As the song says, “whatever happened to old fashioned underwriting?”
knowall, I believe the experienced underwriters who actually thought, underwrote on time tested principles were more accurate than the computer models used by goofy actuaries and modelers.
Yes, that’s why the Aetna, USF&G, and others I can no longer remember were so successful with personal auto.
Long gone since they weren’t thinking or underwriting. USF&G was pretty good at Commercial.
Agent – question for you. (and this isn’t an attack, I really am curious as to what an agent thinks.) 30 years ago when I was a PL Underwriter with a “preferred” company, I was told to never write someone who was divorced in the past year. Now, many companies have abandoned those types of rules in favor of developing a rate(using predictive analytics and additional rating variables)and charging the right rate for such a risk. As an agent, would you rather have the conversation with a recent divorcee about why you cannot write them or tell them you can write them at X price?
CL, thank you for not attacking me and wanting to know how an agent thinks. I have been in this business for over 30 years and have seen every soft and hard market and all the swings up and down. Personally, I have not had any of my preferred markets ever tell me I couldn’t write a divorcee. If it was a current customer, we often had to split up the coverage and write both separate policies. Sometimes, the credit score suffers and a new applicant pays slightly higher premiums. The big problem is when an irate husband or wife tries to take the others car off the policy before the divorce is final. That can be dicey. I would bet a Predictive Modeler couldn’t answer that since they have never dealt with the public before.
Does that explain why our old Standard program always outperformed the great business in our Preferred program?
Really Milner? Our Preferred markets have always done better than the Standard or uprated market. We throw problem accounts to Progressive since they will write about anyone for a price. No need to ruin the good market with substandard business who don’t pay bills on time and have more claims.
Predictive modeling is really a combination of underwriting and ratemaking on steroids. Over the years, many fine underwriters and actuaries noticed trends on general classification like divorce people or young people in sports cars and made rules and rate classes. A sound predictive model can refine those classifications in more detail and constantly update them. A trend that might have been valid 40 years ago may not be valid today.
Have some experience working with ‘friend of the court’ regarding child support calculations — they like to pull out a chart, while sitting in an air conditioned desk environment, and say,” okay your gross income is “x,” and you have 3 kids, therefore you have to pay “y.” Challenged them once about the fairness and rationality of this method and the clerk responded, “take your comments to your lawyer,” and hung up on me.
I guess using these predictive models without any (ongoing) field or underwriting expertise/input could be a mistake.
You are correct knowall. Using the models without any judgment is typically not the recipe for success. Sometimes the models generate indicated paths that cannot be followed due to social, regulatory, competitive and good ole common sense reasons. The best companies know when it is necessary to deviate from the model.
Cl, sometimes the best companies make big mistakes in judgment on their modeling. One of ours did that in 2014 and implemented state wide rate increases, re-tiered their book, re-scored all the customers of their book. Guess who took the heat from all those changes? Net result was that many customers told us to move them or they would find someone else.
You are absolutely correct, Agent. I have definitely seen companies that lose sight of how their decisions impact the action on the street.
It is an interesting question if field or underwriting expertise would optimize the predictive modeling process. I have been in the business for 30 years and wonder if some of my past perceptions may hold me back in this process. In the past we would have special underwriting rules and judgments to compensate for the lack of flexibility in our rating plans and underwriting guidelines. The predictive modeling process, in theory with valid and credible data, should eliminate the bias and indicate the appropriate rate for a risk.
Observer, the key word is “valid”. Often, the modeler data is both flawed and biased. How valid is discriminating against older drivers who are loss free, pay their bills on time and giving them a rate increase for no apparent reason? How valid is Price Optimization schemes designed to raise rates on long term loyal loss free customers. See Allstate for that one and it is happening with other markets as well.
Older drivers have ever-decreasing vision and ever-slowing reaction times. They are more likely to be involved in accidents at 90 than they were at 50.
The problem with your statement is that it usually starts about age 65 with many markets on discrimination by age. Most people are dead by age 90 or haven’t been driving for many years and in a nursing home. Customers we have at that age(65-75) are sharp, have glasses for good vision, don’t have accidents, pay their bills on time and all around good customers who give no one any trouble.
And if you tested those 65-75 year old drivers on how perceptive they are and how quick they are able to react, you would find they don’t have the same clarity of vision or are as fast as reacting than they did at, say, age 35.
This is why older drivers are charged more — they do not see as well as they used to, their reaction times are slower, and those are only going to get worse as they continue to age.
This is just statistically incorrect. “Many,” “some,” “my clients,” etc are irrelevant; the only thing that matters is the population on a whole which makes up the risk pool. This age group is without debate a group that is more dangerous than others and gets increasingly worse every day.
“Mature” drivers can be profitable. However unlike young drivers whose risks characteristics typically improve over time, a senior’s driving skills will diminish over time. Too often, we have senior drivers that will not stop driving until repeated incidents occur and it’s difficult for carriers, from a pricing perspective, to know when this time occurs.
Well said. One up-vote coming your way! :)
You know what Rosenblatt and SWFL, not too many years ago, the market thought mature drivers were a favored class to write. They typically hardly ever have a claim, pay their bills and don’t shop their coverage much. I would much rather have one of these accounts than the 25-35 age class who is so distracted by their cell, they run over someone.
Agent,
In other words, you want the easiest accounts you can get. I wonder if your carriers know this is truly how you think. It may be high time for them to terminate their contract with you and move onto an agent who does not mind putting in the work.
Agent – didn’t you have a personal experience where you had to take your father’s (or grandfather’s, I don’t fully recall) license away? Why did you do that? Did they keep crashing their car or they couldn’t safely drive at night because they couldn’t see so well in the dark anymore? Those are just two reasons why aging drivers see increases in premium even though they didn’t experience a loss — as people get older, they will become increasingly less safe behind the wheel. Do you disagree with me on that?
Ron, the cubicle knat. Had you ever been a real agent, you may have a different attitude about insured’s and the market. I have been putting in the work for a long time, you idiot. By the way, why would the carriers I do business with terminate my contract when I produce consistently profitable business for them and they reward me with excellent contingency each year? We set a new record for profitability for 2015. That is something you will never see. Are you jealous of successful agents or what is it?
Agent,
Congratulations on a record year!! I mean that sincerely.
I also had a record year in salary. Looks like we are both doing quite well in President Obama’s economy.
Does this mean that you will actually pay your taxes this year or do I have to pay them for you like the past several years.
We do well in spite of the Obama economy, not because he has done anything to improve it. We work twice as hard to support the leaches of society. Our carriers do like us, however.
Damn, nice one Ron; although it’s not hard, it’s funny seeing this clown made to look foolish.
He says he prefers the accounts that don’t switch, because the younger people cause more accidents (texting of course), you basically quote him and he denies it. Total clown.
Of course, he’s wrong on the facts too. Younger drivers in the age he cites are slightly more dangerous if you don’t control for urban/rural. But the age he specifically doesn’t want is the exact age they start to become safer and the age he prefers is when they begin to get increasingly more dangerous every year-unsurprisingly. I say unsurprisingly because he had been totally wrong on literally every topic I’ve seen him write about. He cannot agree with data and modeling because he’s so divorced from reality it would challenge every single thing he thinks and he couldn’ttalk handle it.
As for the data modeling, he would be bankrupt almost immediately if he actually had to figure out how much to charge to cover losses, what business to write, etc. Just 100% clueless, and qualified for no other role than that as a salesman where blabbing matters more than actual reality, I don’t believe for a second he could last a month at a legit agency.
Rosenblatt, No, I didn’t have to take either my fathers or mothers car away from them since they died. Some seniors do have a problem at some point seeing. Have you ever heard of a Medical Statement being filled out to list conditions? It is pretty common and starting about age 75, some companies require them. If there are no compelling reasons to list health concerns, it is generally ok to let them drive. By the way, they are far safer than the young cell crowd who are addicted to their phone and often text while driving. I see it every week and it is disgusting.