At a time when we have a president grabbing kids away from their parents and a Supreme Court about to move even further to the right, the impact of auto insurance on inflation may not seem the most pressing matter. But there are aspects to the issue that are informative about how we measure and think about inflation.

First, the importance of auto insurance in overall inflation seems to have gone largely unnoticed. In the last year, it has passed medical care as a driver of inflation. The motor vehicle insurance component of the consumer price index (CPI), which has a weight of 2.4 percent in the overall index, has added more than 0.21 percentage points to the inflation rate over the last year. By contrast, medical care has added just 0.19 percentage points.

The reason for the large impact of the car insurance component is that the CPI shows a 9.0 percent year-over-year rate of inflation in the last year. The rate of inflation in the auto insurance has been rising rapidly since 2013, when it was running at close to a 4.0 percent annual rate.

It is worth noting in this respect that auto insurance is not a major contributor to inflation as measured by the personal consumption expenditure (PCE) deflator. In the year from the first quarter of 2017 to the first quarter of 2018, it added just 0.04 percentage points to the rate of inflation, as measured by the PCE.

Part of this difference is because the PCE measure showed a somewhat lower 7.5 percent rate of inflation over this period. (The PCE measure showed just a 4.9 percent inflation rate in auto insurance in the prior year.) However, the more important factor is the much lower weight of auto insurance in the PCE deflator. Its weight in the PCE of 0.54 percent is less than a quarter of the weight in the CPI.[1]

The reason for this difference in weights is that the PCE measure is using the net cost of auto insurance, whereas the CPI measure is the gross cost. The net measure is the cost of insurance after deducting the claims insurers paid over the course of the year. The gross measure just picks up what households pay for their policy.

For example, if a person pays $1,000 a year for insurance on their car, this would be the gross cost of insurance picked in the CPI. A 9.0 percent increase in the price of insurance shown in the CPI would mean that the cost of this policy had risen 9.0 percent to $1,090.

By contrast, the net measure would be subtracting what an average policy paid out in claims over the course of the year. If this came to $800, then the net measure would be recording this policy as having a net cost of $200. The 7.5 percent increase reported in the PCE would imply an increase in the net cost of auto insurance to $215. This could be due, in principle, to either a rise in the gross cost of the policy of $15, a reduction in the average payout of claims of $15, or more likely, a mix of the two.

To make matters slightly more complicated, neither of these measures correspond to what people actually pay for insurance. The Insurance Information Institute (III) reports that the price of an average policy increased at just a 2.8 percent annual rate from 2011 to 2015, the most recent year for which data is available. While there may be some issues about the accuracy of the III’s data, it is also, in principle, measuring something different from the government price indices. It measures the average cost of an insurance policy. By contrast, the price indices are looking at the rate of inflation in a quality-adjusted measure.

If the average person is able to prevent their premium from rising by having a larger deductible or smaller cap on claims, this would not translate into higher prices in the III’s data. However, it would translate into a price increase in the CPI, and most likely the PCE as well, since people would be paying more money for a policy that provided the same benefit.

Moving back to the net versus gross question, if we think that people view insurance as a cost, without regard to what they get back in claims (which for most people in most years is zero), then the gross measure in the CPI probably is what they would care about in assessing their cost of living. Even if they do get paid on a claim, this is the result of actual damage they have incurred.

So if they run into another car and do $2,000 of damage, they can collect $2,000 from the insurance company, but this money must be paid to the other driver and the repairs on their own car (ignoring issues of deductions or improper charges). The claims offset a loss, they are not an addition to income.

This problem comes up in an even more extreme form in health care insurance. For this item, both the CPI and PCE use a net insurance measure. However, the vast majority of people get considerably less back on claims than what they pay in premiums, since most people are relatively healthy.  

This creates a somewhat peculiar situation in the way costs get picked up by government measures. Suppose that the government requires insurers to cover everyone without regard to preexisting conditions, as was done under the Affordable Care Act. This will lead to a rise in the cost of the average policy since insurers will now be covering people with more health issues who have higher costs on average.

However, this does not show up as an increase in the cost of health care insurance in the CPI, since the increase in price is actually associated with an increase in the average amount of services being provided. In effect, healthier people are subsidizing people with health issues through the price of their insurance. But, this cost does not show up as a rise in the cost of living, since the higher cost to healthier people is offset by the lower cost to less healthy people.

The policies of the Trump administration have been largely designed to undo this subsidy as they make it easier for healthier people to opt out of insurance pools that include people with health problems. This has resulted in sharp increases in the price of health insurance in the individual market.

While the Republicans have claimed that they do not want to end the ban on discrimination against people with health conditions, this has been the effect of their policies. It is now far more expensive for people with problems to purchase insurance.

Returning to car insurance, while the cost has been rising rapidly in the CPI, it does not appear to be resulting in a larger share of expenditures going to auto insurance. According to the Bureau of Labor Statistics’ Consumer Expenditure Survey, the share of spending going to auto insurance remained constant at 2.0 percent between 2012 and 2016.

This means that households have largely been able to offset the higher price of auto insurance by buying lower cost policies. That could mean that they are more exposed to expenses in the event of an accident, but the amount of money that goes out of their pocket for insurance is not rising as fast the auto insurance component of the CPI indicates.

It is likely that households really do see the move to less comprehensive auto insurance as a deterioration in their standard of living, just as if they had to get health insurance with higher deductibles and co-pays. The cost of living notion is a difficult concept to pin down precisely and insurance is a case where it is especially hard, as indicated by the difference in the treatment in the CPI and the PCE.


[1] These data can be found in the National Income and Product Accounts, Table 2.4.5U, Line 273 and Line 1.
Vote on CREDO