A View From The Press Box

It is well established in research that driving behavior changes somewhat when gas prices increase, and that these behavioral changes lead to a decline in number of automobile insurance claims that are filed.

Consumers simply respond to economic incentives and disincentives by reducing their driving as gas prices rise. As a result, auto insurance claim frequencies drop. Consequently, higher gas prices lead to lower auto insurance claim costs.

Mr. Leroy Boison, a Fellow of the Casualty Actuarial Society, studied this effect in 2005. (“Will Post-Katrina Gas Shortages Impact Auto Claim Frequencies?,” Pinnacle Actuarial Resources, Inc., December 2005.)

After studying the 1979-1980 energy crisis, Mr. Boison found that: “The crisis did not just contribute to short-term reductions in auto claims frequency; it also contributed to a long-term decline. For several years after the political crisis had passed and gasoline prices declined from their historic highs, claims frequency failed to return to pre-crisis levels.”

One of the reasons for this development, he noted, was that “…drivers found other ways to get around and stuck with them. These other factors include carpooling, public transportation and consolidating errands, which reduced drivers’ exposure to the perils of the road.”

Mr. Boison concluded that: “Since increases in gasoline prices contributed to a long-term decline in claims frequency as drivers opted to put fewer miles on the road, it is reasonable to assume the same could have happened after this crisis if the increase in gas prices was significant and remained at high level.” Other studies have also confirmed this effect.

Auto insurance rates are based on the trends regarding the cost and frequency of claims. The cost of claims is not affected by gas prices, except very indirectly.

However, if the cost of gasoline makes people change their driving behavior, then the frequency of accidents and, thus, claims, is directly affected.

Using the most common actuarial method, we found that the annual rates of change in claim frequency indicated by Fast Track data through September 2007 are:

Bodily Injury -5.2%
Property Damage -2.4%
Comprehensive -8.2%
Collision -1.8%

In other words, claim frequencies were dropping by significant percentages per year nationally even before the recent skyrocketing gasoline prices. These trend factors must be altered downward to reflect current conditions and the mounting evidence of significant changes in the driving behavior of Americans.

Federal Highway Administration data shows that passenger automobile miles driven dropped by 0.7 percent from 2005 to 2006, well before the current spike in prices. In March, the U.S Department of Transportation reported that Americans drove 11 billion fewer miles than the same month a year ago, “a month-on month percentage decline (which) is the largest since record keeping began in 1942.”

ALL GRADES
NOMINAL REAL

1998 $1.12 $1.42
1999 $1.22 $1.52
2000 $1.56 $1.88
2001 $1.53 $1.79
2002 $1.44 $1.66
2003 $1.64 $1.85
2004 $1.92 $2.11
2005 $2.34 $2.48
2006 $2.64 $2.71
2007 $2.85 $2.85

The current price of gasoline, over $3.50, is at an all-time high in inflation-adjusted terms. It is incumbent upon each state’s insurance department to test the current pricing structures in auto insurance to determine if the changes in driving behavior are at least partially to blame for the remarkable profits of property-casualty insurers in recent years.

They should also evaluate whether current extreme gas prices will lead to further windfall profits for insurers as drivers alter their driving habits to ease the pain.

It would be a mistake if, in performing this analysis, regulators relied only on long-term trend factors without adjusting claim frequency expectations downward to reflect the driving behavior changes induced by the sharp gas price increases of the last few months.

From a consumer standpoint, the annual survey I complete for my auto insurer includes a section on annual miles driven, and each vehicle is rated on that estimated mileage.

The cost differentials between a vehicle driven 0-5,000, 5,000-10,000 and 10,000-15,000 annual miles is very significant. The explanation is that there is additional risk placed on the company due to the exposure you have while driving those miles. I don't have any argument with that at all.

However, if you do not drive anywhere near those miles at the end of the policy year, do you get a rebate on what constitutes an overpayment on estimated premium? No! If you request one, it's just laughed off as some sort of joke. The company now has had much less exposure while you paid for the higher levels.

To add insult to injury, for many years, if you had two drivers and three cars and sold one, some insurers would add the mileage driven annually for the sold vehicle to the highest mileage rated vehicle, or split that mileage between the other two, stating that in their mind it was logical if you drove "X" miles total, you would still be driving "X" miles but dividing it by two rather than three.

That happened to us on several occasions when we sold specialty vehicles that we drove only occasionally for vacation trips, that we would not have taken with another vehicle, nor used as daily drivers.

What the current system does is create a whole system of fibbers who underestimate their mileage each year and then have to update it, if and when necessary. Or, people who create extra paperwork by changing the mileage data policy mid-year by calling the company with an updated mileage figure, up or down. Or, the unlucky ones who have an accident and get caught with the lying face of the odometer in full view...or is that the truthful face of the odometer and the client...

So, yes, if insureds are driving less, cut the rates based on driving mileage IF the insurer is using mileage to distill the current rates. Otherwise quit using mileage at all and use driving records and other quantifiable data as the source of premium charges.

In review of Mr Hunter's comments above, could one not then claim that the vehicles with the higher miles-per-gallon rating should pay higher premiums since they may not change their driving habits as much as someone driving a lower MPG SUV?

The higher fuel prices have a lot less impact on someone driving a small car getting 42 MPG vs. someone driving a larger SUV that gets 15 MPG.

In that same example, an identical accident with the two different vehicles would have different exposures for both physical damages/property damages and for medical costs for the occupants involved.

The smaller car would probably have higher physical damages and they protect the occupants less, so the injuries to the driver and passangers could be greater. Should the drivers of the smaller cars with better fuel economy get surcharged for these reasons?

Since we are trying here (and the New York superintendent is trying) to draw conclusions on what we THINK will happen in the future, two things come to my mind.

First, how different is this from homeowners carriers trying to use predictive catastrophe modeling and pricing homeowners coverage based on what they THINK will happen? They are only just starting to see if they think some of that should be allowed, why is it that they think this sort of 'what if' thinking should be immediately applied to car insurance?

Second, insurance carriers have spent a number of years now promoting insurance scoring as the most predictive tool they have for setting auto rates. I think it's safe to say that, with all the economic problems, the mortgage crisis, increased fuel and food costs, etc., the average insurance score for the general population will be going DOWN, not up.

So under the same logic, shouldn't the companies be raising rates based on the fact that scores are dropping, which theoretically will lead to higher claims costs?

Driving falls but premiums rise.

Since the private passenger auto exposure unit is the car-year, it is likely in many circumstances for driving miles to fall, while at the same time claims-per-100-car-years rise.

In fact, Bill Martin of Farmers Insurance pointed to this likelihood in a Reuters article last month. He was quoted as saying that “many consumers are choosing to not buy insurance as they give up their car or find they can no longer afford it when choosing between long-term financial security and gas or food. We may see insurance prices increase in spite of lower mileage per driver."

Mr. Martin subsequently confirmed to me that the following circumstance conforms to what he meant: If a household reduces its insured cars more than its drivers reduce their insured driving, the average miles per-car rises and with it the per-car cost transferred to the insurer. If this happens broadly, eventually insurance prices must also rise.

This scenario also explains why lower credit scores correlate with more claims per-100 car-years (including Uninsured Motorist claims).

Households under financial stress must cut expenditures of all kinds. In general they cut miles of driving, but usually they cut their insured cars proportionately more than their miles. Hence miles per car rise.

Of course, number of miles driven would not affect insurance prices if the exposure unit were changed from car-year to odometer-mile. Full information on the two exposure units and their effects is available at www.centspermilenow.org .