Car Insurance Is Too High and Healthcare Insurance Not High Enough

The Days Ahead:

• Short week. Jobs numbers

This Week:

• The Fed’s view of inflation is…different.
• And they’re focused on some very strange line items right now
• The auto insurance and repair market is going through some big changes
• Premiums may stay high
• The way the BLS measures health care insurance is very strange
• The end of LIBOR as we know it
• Halfway through the year and capital markets are a lot better than 2022

What the Fed Tracks

The Federal Reserve (Fed) has talked about more than one inflation measure in the last year or so. It’s go-to measure is the Personal Consumption Expenditure price index or PCE, which differs from the headline Consumer Price Index (CPI). The first measures what businesses are selling and the second what consumers are buying. This seems odd. Surely, they’re the same in the end?

Not necessarily. A quick way to think of it is how consumers can change buying habits. Say, bread and corn flour are 1% of both indexes. Let’s assume bread prices go up. They will immediately drive up the CPI. But if bread prices go up a lot, consumers may switch to corn flour wraps because they’re cheaper (this is called substitution but that’s a long word for switch). Businesses will then sell more corn flour.

In the PCE, that will show up as a halving of the bread index and a doubling of the corn flour index. The weighting in the CPI won’t change. I mean, it might eventually if consumers permanently switch from bread to corn flour but they won’t change the index immediately. But because the PCE measure what businesses are selling, the change to corn flour will be picked up fairly quickly.

Anyway, in this example, the CPI will go up more than the PCE because it’s sticking with the old 1% weight of bread. A good rule of thumb is that CPI will be higher and more volatile than PCE. And here it is.

Over the last 63 years, the CPI ends up 32% higher than the PCE. Since the beginning of 2022, when the Fed started to talk seriously about inflation, CPI has risen around 8% and the PCE by around 6%.

So far so fine. The Fed then started about six months ago to talk about “core-core” services inflation which is services, less housing and energy and is about 25% of the total CPI. The Fed still talks about PCE, as here last week when Chair Powell gave his Congressional report, but he quickly changed to the “core-core” inflation. Why? Because he’s more concerned about wages and believes the wages will show up in the “core-core” services because labor accounts for the biggest cost for most services.

Again, that’s fine. But the remaining “core-core” is overwhelmingly driven by five components: hotels, health insurance, airlines fares, auto insurance and repairs. We’ll dig into two of those to support our point that wider inflation is easing.

Check Up on Auto Insurance Prices

The latest inflation report contained good news with the annual rate in May at 4.0% against 4.9% for April. The “core” which excludes energy and food was at 5.3% as against 5.5% in April, with transportation services, which includes things like used cars, motor vehicle maintenance, and insurance all up 12% to 17% over the year. Together they make up about 16% of the core index, so their price moves count.

The increase in vehicle maintenance and insurance, of course, go hand in hand. If you’ve noticed your car insurance premiums have gone up it’s down to a number of trends which don’t bode well.

One is a shortage of qualified mechanics, many of whom were laid off for months during Covid-19 and either waited to return, retrained or retired. The Bureau of Labor Services (BLS) puts the number of auto service technicians at 664,000 down from 733,000 in 2020. It’s a profession which has around 73,000 openings a year despite a projection of zero growth for the next 10 years. It’s not a profession that’s growing but it remains in high demand.

And as for car insurance, it’s up 17% on the year and up 47% since the Covid-19 lows. It’s run much faster than general inflation.

It’s not as if insurance companies are raking it. Allstate has underperformed the S&P 500 by 30% this year and Progressive by 12%. In a recent earnings call, the CEO of Progressive mentioned:

…higher-than-expected severity trends in previously closed claims in personal auto, primarily in fixing vehicle coverages”

The reasons were not enough shop capacity, higher labor, parts and repair costs.

Allstate was more forthcoming than Progressive and showed the combined ratio (CR) of its auto insurance division. Normally a CR is the sum of losses divided by expenses. The lower the number the better. If I collect premiums of $100 and pay claims of $95 my CR is 95 and all is good with the shareholders, board and actuaries. Allstate’s CR is usually 92 to 96 but last year went to 110, meaning it collected $100 in premium and paid claims of $110. No one is happy with a CR of $110. Also, the level of catastrophic loses, where there is no residual value of the vehicle after an accident, jumped from 0.9% of claims to 6.5%.

Again, it seemed to be because of more severe accidents, higher repair costs, medical inflation and “attorney representation,” which is a new term for us but suggests drivers were lawyering up and ready to litigate.

So, the higher premiums seem to be about high repair, health care and parts costs combined with more severe accidents and a shortage of people to mend the cars.

It may also have something to do with driver inattention, watching very big screens in their vehicles. Some screens are up to 17”, which was a good sized TV a few years ago. Some stretch across the drivers’ dash, measuring over 50”.

A recent study showed that it took a BMW iX driver 30 seconds to complete four simple tasks like changing a song  or adjusting the climate control, during which time it traveled 800 meters. One hopes the tasks were done with the occasional look up at the road. In a Tesla, the tasks took 23 seconds and 700 meters and in a 2005 Volvo, with no touchscreen of course and just a few dials and buttons, it took 10 seconds and 300 meters.

We’ve also mentioned here that certain electric cars burn a lot. If the lithium cells catch fire, it can take 6,000 gallons of water and many hours to extinguish. And after hours on fire, the end damage will probably not buff out. There’s a related problem that the same certain electric car uses batteries for its subframe so a side-on hit will bend not just a couple of metal bars but a $20,000 sophisticated power plant that costs more than the remaining value of the car. That would go some way to explain Allstate’s higher “catastrophic” losses.

Electric vehicles (EV) also weigh more than conventional cars. A popular EV model, made not 100 miles away from my desk, thumps in at 5,100 lbs of which the battery weighs 2,000 lbs. The BMW i7, which ate all the pies, is 5,900 lbs which is not far off the basic weight of a Cessna Citation jet. A Toyota Corolla comes in at 2,900 lbs. Now, as the Institute for Highway Safety points out:

Heavier vehicles also tend to continue moving forward in crashes with lighter vehicles and other obstacles, so the people inside them are subject to less force.”

Which is good if you’re the heavier vehicle but not so good if you’re in the Toyota. 

Some people therefore may say that electric vehicles cause insurance rates to increase. Others that all the self-driving aids make them safer and cheaper. We won’t say either because some people are very litigious. We do know that car insurance rates are going up through a convergence of a several one-off adjustments. They will eventually stop rising by so much but for now they stand out as an increasingly expensive household item.

Check Up on Health Insurance Prices

The recent inflation report also had an update on health insurance which fell 3.7% over the month and 20% over the year.

That prompted the following from Bloomberg:

…there’s one startlingly positive news item buried in the data, it concerns the highly politicized but extremely important business of health insurance… The latest data suggest that premiums are down more than 20% over the last 12 months. That might just be a handy spur to consumers’ disposable income if it continues”.

No, no, that’s not how it works. We wrote about this last October in some detail but here’s a recap.

The health insurance numbers in the CPI do not measure the price of your healthcare a year ago to the one you just paid last month. The BLS gave up trying to calculate health insurance some years ago because, in their words, they could not control for quality or risks. When you dig into an insurance premium and figure out coverage, networks, allowed procedures, deductibles, co-pays, maximum out of pockets, and family plans, you soon hit a brick wall trying to compare one plan to another, yet alone the costs changes over a year. If I pay $500 a month for my Kaiser plan and last year paid $600 for my Cigna plan, that may look like a cost reduction. But if my co-pay, deductibles and so on all went up, then my “savings” may not be anything more than a cost shifting.

So the BLS uses something called the “retained earnings model” where they look at all the premiums collected by health insurance companies and subtract all the benefits paid. If that sounds like a fiendishly difficult exercise, you’re right, it is. Here it is. The data is from the National Association of Insurance Commissioners or NAIC and it’s, mostly, the data used by the BLS.

There are two things that matter.

One, the NAIC publishes the data once a year in October with a 10-month lag. In October 2022, we saw the numbers from year-end 2021 and they compared the numbers to 2020. So, when we look at the inflation numbers in the CPI report in 2023, we’re looking at the changes from 2020 to 2021. Yes, two years old.

Two, the retained earnings model will show prices going up if insurance companies had a good year and down if they had a bad year. In 2020, health insurance companies took in $834 billion in premiums and paid out $683 billion in claims or benefits (they’re synonymous in the health care world). The difference, or retained earnings, was $151 billion. In 2021, they took in premiums of $900 billion, up 8%, and paid out claims of $775 billion, up 21%, for a difference of $125 billion. By this method, the retained earnings fell 18%. That’s what’s showing up in the CPI (the BLS actually do one more calculation but it doesn’t change the outcome for 2021).

If you remember 2020, people delayed all sorts of non-urgent medical care so health care companies saw a big drop in claims. Fast forward to 2021 and people felt safer about seeking healthcare and started to use services more. Demand jumped. That depressed health care company earnings and, again, shows up as lower earnings and so a drop in healthcare inflation.

The BLS knew back in October what the change in earnings was, so what they do for the next 12 months is smooth the earnings. We knew coming into 2023 that health insurance premiums in the CPI world would fall around 3% a month. And that’s exactly what’s happening and will do so until October this year.

As we said before, these numbers can do you your head in and we sympathize with the BLS in trying to make sense of them. What we do know for certain, is that healthcare premiums are not going down and will not go down!

We can get a better bead on the way health care costs are headed by looking at producer prices. They show up in producer prices because most healthcare premiums are paid through group employer plans. So, it’s a cost for producers more than individuals. Here’s how it looks.

The blue line is the CPI version of health care, showing a fall in prices that none of us will see. The green line is what employers pay and it has the steady, relentless increases we all actually experience. Note the bump every January when new group plans come into effect. Since 2006, general prices are up 53% and group health insurance up 81%. Our guess is that the CPI health insurance component will run next year at about the same level as the PPI price, which is about 4%.

Normally, we wouldn’t dive into two relatively small components of the CPI in such detail but the Fed made us do it with their “core-core” focus. If we strip out the five fast moving components of the core-core, we’re left with an inflation rate of around 3.5% and trending down. We know that the auto insurance/repair price index will settle eventually, albeit at much higher rates than we’re used to. Healthcare will revert to far less volatile changes. And when those kick in, we should be looking at a much calmer inflation outlook.

Goodbye LIBOR, We Won’t Miss You

In 1969, Minos Zombanakis had a problem. He worked at Manufacturers Hanover Trust (now J.P. Morgan) in London. He had a willing customer, the Iranian government, who wanted to borrow $80 million from London banks. They also wanted a variable rate loan because they were selling oil and felt that was as good a hedge as any in case currencies moved against them. But no single bank was willing to lend all that so Minos called around a few banks to organize a syndicate. So far so good, but what rate would they offer?

He came up with a great idea that he would survey each bank and they would take their funding rate, which was the cost of funds, add a bit for profit and report that rate into a pool. The average of the rates was then used as the cost for the Iranian loan. He called it the London (because it was for banks in London using offshore dollars) inter-bank (because it was between the banks) offered rate (because it was the rate the banks were prepared to offer to clients). And LIBOR was born.

Soon the rate drove all syndicated bank loan markets, then all floating-rates notes and bonds and then mortgages. Then in 1997, the Chicago Mercantile Exchange (CME) launched Eurodollar futures so that companies could borrow at fixed rates in the future. They decided to use LIBOR as the benchmark rate. Other currencies joined in and so they added term loans going out three months to five years. You can see where this is going. By around 2010, the banking and futures industry had hundreds of trillions of dollars agreements, all pegged to many versions of LIBOR.

During all this time, LIBOR was still a survey. It wasn’t a result of actual price action, like a bond or a loan. The survey also allowed banks to use any number they wanted and could submit the lowest or highest bid or any other “arbitrary selection.”

Well, if you think that had the makings of some shenanigans you’re right. Around 2010, the banks had two incentives to report either a higher rate or lower rate.

They could, and did, report a higher rate because banks owned futures desks and if a futures contract was about to roll over, they could report a higher number than their actual costs and make more money on the contract.

They could also, and did, report lower numbers because a higher number would show that the bank was having trouble accessing funds and customers may conclude it was in trouble. By reporting a lower LIBOR it looked as if the bank was in better shape than it was.

Well anyway this all blew up with embarrassing taped conversations and charges of manipulation. Banks were fined around $9 billion and 38 traders charged.

Everyone agreed LIBOR was bad but what was the alternative? Some, like the European Union cleaned up the process and continue to use Euribor contracts. But U.S. banks needed a new benchmark. LIBOR was given a few years to wind down and everyone looked for an alternative. LIBOR closes at midnight on June 30, 2023. So, what’s the new LIBOR?

The Fed came up with the Secured Overnight Funding Rate or SOFR. It’s the rate at which the Fed will lend money to you if you show up with some Treasuries and need to borrow money overnight.

There are two problems with that.

One, banks lend for longer than overnight so they would have to add on a term, or time, premium to SOFR.

Two, it doesn’t reflect banks’ cost of funding. You could argue that neither did LIBOR if they were submitting made-up numbers but, in this case, what the Fed lends at is not what banks can borrow at.

The Fed has run SOFR for a few years now, just so everyone can get used to how it works. Here’s an overlay of SOFR and the soon-to-be-retired LIBOR.

SOFR is running about 0.15% below LIBOR, which makes sense as, again, it’s the Fed and Treasuries, not banks and commercial loans.

It doesn’t mean loans will be cheaper but the benchmark rate should be less volatile. Ultimately, the SOFR rate is pegged to things like the fed funds and various other money market rates. The Fed won’t want to see those swing around in the way LIBOR used to. It could get weird at times like in early 2020 when SOFR rates collapsed to well below market rates before the Fed could reset policy rates. And there could be problems when bank creditworthiness matters and SOFR remains unchanged because it has no credit component…. it’s just a bank, any bank, showing up with Treasuries at the Fed. The Fed asks no questions about credit.

Anyway, this could be a big fat nothing-burger or we could see some adjustments as markets and banks adjust to the new regime. But come midnight Friday, LIBOR will only be in the history books.

The Bottom Line

Time flies pretty quickly these days. Either that or I’m just getting older. But here we are at the halfway mark for the year and the S&P 500 is up 14%, the Nasdaq up 29%, European stocks up 14% and Japan up 27%, although that last number trims to 16% for a U.S. dollar investor. Stock volatility, as measured by the VIx, which measures stock volatility, has rarely moved up above 20, except for the bank scare in March, and its bond counterpart, the MOVE, are around 18-month lows. Oil is back to its pre-Ukraine invasion days and long-term U.S. Treasuries are unchanged. Heck, even the extraordinary events in Russia, an invasion, an attack, a coup attempt, a reckoning, a bargain and an exile all in 48 hours, failed to startle markets.

Yet, we’ve had another 0.75% of tightening from the Fed on top of last year’s increases of 4.25%. The economy, while weaker in some areas like housing, is holding up well and employment has shown steady, if not increasing gains.

We showed this chart at the end of last year to illustrate just how extreme the stock and bond moves in 2022 were. We’ve updated it to show where we are so far in 2023.

As we mentioned, we’ve only seen four years since 1929 when both stocks and U.S. Treasuries were both down and the worst was 2022. In hindsight an inflationary surge from a zero level of interest rates was asking for trouble. The good news is that we’re not at the zero bound anymore and we don’t see any excess or leverage in the broad economy.

We expect some calm in the market for the next few weeks. The July 4^th holiday should keep nervous traders away from their desks at least until the next Fed meeting on July 26^th and the earnings season kick off around July 14^th. Have a great July 4^th week everyone. No blog next week and apologies for not getting one out last week (don’t ask!)

Subscribe below for our investment updates.

---

Art: Petra Cortright (b 1986)