US CPI inflation overestimating auto insurance prices, explaining 1/3 of the gap with PCE

Headline US CPI motor vehicle inflation in March reached its highest level since 1976. However, official estimates of personal auto insurance inflation vary widely, ranging from 6.5% to 22.2%. Our own estimate sees motor vehicle insurance inflation at 14%. The possible overestimation in official measures may distort the disinflation signal and will continue to challenge policymakers.

• Headline CPI motor vehicle insurance inflation has reached highs not seen since 1976, but estimates have also started to diverge widely.
• Headline CPI reported year-on-year auto inflation at 22.2% in March.
• Based on industry rate filings, we estimate annual motor insurance inflation at 14%, suggesting the headline CPI motor insurance print overestimates by 8 percentage points.
• The CPI over-estimate explains one-third of the 100 bp gap between core CPI and core PCE.
• This is another reason why the Fed prefers to target PCE inflation.

In March this year, official headline US motor vehicle consumer price index (CPI) inflation was reported to be 22.2%, its highest level since 1976. However, in our view the headline print may overstate the premium price rise in personal auto insurance (see Figure 1). Based on industry rate filings, we estimate that premiums were up 14% year-on-year in March. We also note wide variation in other official estimates of personal auto insurance inflation. For example, the producer price index (PPI) series¹, put motor inflation in March at 8.6% (with investment income) and 6.5% (without). The discrepancies in official data could slow reported disinflation and complicate monetary policy decisions.

Figure 1: CPI "overstatement" due to motor vehicle insurance (percentage points)

The premiums households pay for their car insurance have risen steeply over the last two years, as insurers have sought to repair heavy underwriting losses.² In recent months, auto insurance CPI has surged and diverged farther from comparable inflation series than ever before (with the exception of a short spell during the pandemic when different treatment of discounts/rebates yielded a large gap). There are four official series for US private passenger auto insurance inflation: motor vehicle insurance CPI, two PPI series,³ and the personal consumption expenditures (PCE) deflator for net motor vehicle & other transportation insurance (based largely on PPI with investment return).

The gaps in inflation estimates could be due to differences in methodology. For example (1) CPI is based on urban households; the PPI scope is broader;(2) the main PPI series considers investment earnings; CPI does not; (3) CPI measures what households pay and PPI what insurers receive; (4) CPI follows a sample of policies regardless of policyholder behavior and does not account for consumers switching to cheaper policies; PPI only follows a policy as long as it is in effect; (5) the weight applied to auto insurance in the PCE is based on premiums less claims; the CPI weight is based on full premium amount.

The first three reasons do not explain much of the difference – over 90% of households are urban, the gap between the PPI series with and without investment income is small, and factors that can create a gap between consumer payments and carrier receipts (eg, changes in premium taxes or broker commissions) are limited. The fourth reason likely matters more. CPI and PPI started to diverge right at around the time that consumers started shopping around and policy retentions started to drop, but that still does not explain the whole increase. Other factors lacking transparency are likely at play too. We hence look at insurance sector data for alternative benchmarks. 

Based on industry revenues and proxies for exposure growth – personal auto insurance premiums written were up 17% year-on-year in 4Q23, while US real GDP grew by 3.4% in 4Q23 and new light vehicle sales by 3.7% in 1Q24. We view the March CPI (22.2%) as a potential overestimate and PPI (6.5%) as a potential underestimate of auto insurance inflation. Our proposed alternative measure is based on industry rate filings,⁴ comparing approved and filed premium increases with total personal auto earned premiums. This derives our estimate of March personal auto insurance inflation of 14% on a trailing 12-months basis (see Figure 2).

Figure 2: Private passenger auto inflation measures, %

This data ambiguity is one reason why the Federal Reserve prefers to target core PCE inflation, which as of March was 98 basis points lower in annual terms than core CPI inflation. Among other differences, motor insurance is just 0.6% of the PCE basket, four times lower than the 2.8% weight in the CPI. This accounts for the more significant CPI over- than PCE underestimation. We estimate that more than 30 basis points of the gap between core CPI and PCE can thus be attributed to data distortions in motor vehicle inflation.

The different measures of motor insurance prices complicate the outlook for monetary policy. The Fed can look through elevated insurance inflation due to its lagging/distorted nature, but markets react to monthly and annual CPI prints. We expect that sticky motor insurance prices in CPI will likely continue to contribute to an elevated rate of core CPI inflation into 2025 and distort the disinflation process.