With Fed Chair Powell changing his tune about “transitory” inflation, I thought it would be worthwhile to begin a deep dive on the challenges inflation brings to insurers because there appears to be a lot of false confidence out there. I guess that makes insurance companies more like the Fed Chair than they realize?
We’ll start with personal lines because the feedback mechanism is more immediate and it’s easier to show where the consensus is askew. To jump ahead to the punchline, there is a difference between expected and unexpected inflation and insurers appear to be more focused on inflation expectations rather than the potential for shocks.
Chair Powell has belatedly begun to realize this about the economy in general and eventually the insurers will come to the same conclusion about inflationary pressures on claims costs.
Managing Expectations
As mentioned, there are two types of inflation. No, not transitory vs. recurring. There is expected and unexpected inflation.
“Expected” inflation is something you can budget for. Think of health insurance costs. You may not be able to predict exactly how much they will go up next year, but you can expect it to be within a range and make plans accordingly.
The killer is unexpected inflation or, more simply, surprise inflation. This, by definition, isn’t priced for. However, this is only a small part of the problem.
If there is a one time spike, well, you get behind the curve, but you can quickly raise pricing for it and move forward and keep the harm limited. Most insurers in their public commentary seem to think this is what they’re facing currently.
The problem is when unexpected inflation becomes like a game of whac-a-mole due to broader economic disturbances. New areas of inflation keep popping up due to continuing supply disruptions. The ability to anticipate the next disruption is difficult.
Pricing Implications
When you have unexpected inflation, you are by definition behind the curve on pricing. Worse, you don’t really know if pricing is needed. Should we react to the rising price of lumber or assume it will go right back down? Do we react to the shortage of used cars and assume our cost for a total is higher and thus raise prices or do we assume it’s “transitory”?
It might be easy to suggest insurers will err on the side of raising prices, but what if inflation is transitory? Now, you’ve unnecessarily raised prices and you’ll start losing share. Not all companies will decide to be cautious. Some will choose to maintain volume and gamble on the persistence of inflation.
And if some choose to sacrifice pricing, then those that raise prices will necessarily sacrifice share. And we know how long that tends to last, especially in personal lines.
But even if everyone remains disciplined, how will regulators react to pricing for unexpected inflation? We’ve already seen Texas shoot down several auto rate filings that anticipated rebounds in frequency from COVID recovery. Why would the response be any different for anticipated severity spikes?
By definition, planning for unexpected inflation is anticipating severity. Regulators may not let you price for the increase until after it happens. Which means you are continually behind the curve and chasing your tail.
Play out the scenarios on this and it becomes clear that insurers will fall behind on price. You can’t proactively predict increased severity because you won’t be allowed to take it. When you actually see inflation, you can raise price for it if you expect it will continue, but if you’re right, all you’ve done is tread water and if you’re wrong, you’ve overreacted and will lose share so will likely unwind the rate in your next filing.
It is an environment where it is a lot easier to make mistakes than to get things right.
Severity Implications
There is another issue here that deserves more attention. Inflation isn’t simply about how quickly can we recognize it and price for it. It can also change claims patterns.
The most obvious of these is demand surge. When inflation is driven by shortages, time (and cost) to repair goes up due to lack of supply of materials or labor. Recent events have made clear that demand surge is exacerbated by supply shortages.
However, there are other ways in which similar dynamics can play out. Think about auto accidents. There has been an increasing trend over the years to totaling cars rather than repairing them. This was partly a function of the increased complexity of repairs as well as the price of used cars.
With the shortage of used cars, the balance has now swung away from totals and more towards repairs. That is a meaningful change. However, the cost of repairs is also going up due to the lack of supply of parts and pressure on labor availability, especially as body shop utilization goes up due to increased need for repairs.
This will also pressure retention potentially. Think about it. Body shops take longer to turn around repairs, customers get mad, and hold the bad claims experience against the insurer. There goes the NPS!
So insurers are facing pressure on severity leading to higher loss ratios, while also having to decide if this effect will be recurring or temporary. If the bulge in used car prices ends up like lumber where it wasn’t sustained, then raising premiums to assume 25% higher replacement values is going to cause a lot of shopping. What to do???
Oh, there’s one more problem. All the investments auto insurers have made in UBI? That’s nice and all but UBI only predicts frequency, not severity. If severity becomes a bigger driver of losses than frequency, UBI’s value is diminished.
Consumer Response
The other consideration is consumer behavior. There is the obvious issue of how consumers will react to price volatility. We know they will shop more.
But there are other issues beyond that which will likely lead to confusion and possibly bad outcomes. Let’s revisit that issue with the auto accident.
That’s one that may work in the customer’s favor. Most drivers would prefer their car to be repaired than totaled, particularly when used car prices are low and they may end up underwater on their loan.
When there is a need for a total, drivers will now get higher payouts (though that’s not much consolation given they have to pay more for a replacement). One can imagine drivers affected by the storm related floods over the summer don’t mind that their car was totaled rather than repaired.
Where it can work against the buyer is something like home repair costs. Because your home policy sets your insured value, any unexpected inflation before your next renewal increases the risk of not having enough coverage in the case of a major loss.
The lumber volatility is a good example of the risk here. If your renewal was nine months ago and you insured your home for say $400,000, if your home burned down at the peak of lumber prices (and higher labor costs), maybe it would cost $500,000 to rebuild and your insurance payout wouldn’t be enough to rebuild.
It also is unlikely insurers were doing accurate mark to market updating of lumber prices in their renewal notices. Even if they had accurate numbers, they likely wouldn’t have wanted to disrupt the market by doing real time pricing and disadvantaging those with renewal dates that overlapped peak market volatility.
No Relevant Experience
Probably the biggest risk for insurers is the lack of experience with this environment. It is approaching 50 years since the last bout of supply driven inflation. No actuaries employed today have the scars to know how to respond to data that doesn’t fit what they have seen before.
In fact, we know most companies will fall back to what they know which is the last say twenty years of data and try to fit the new trend to old trends in their toolbox. This won’t end well.
The insurance industry is historically bad at reacting to new trends that don’t accommodate the actuarial department’s muscle memory. It is why people like me can anticipate inflection points better than the actual practitioners! I can look at things without having to anchor whereas companies tend to be very hesitant to make a bold call that breaks with recent trends.
Managements will have to somehow convince employees used to not stirring things up or putting their necks out to be more proactive and weight new data points more heavily than they typically would. This won’t be easy.
It will also take a different skill set. Insurers may need to hire macro analysts from outside the industry to proactively identify changes before they show up in loss data rather than wait for adverse trend and search afterwards for an explanation (though this brings us back to the earlier challenge around convincing regulators to let you price on theories and not data).
If this reminds you a little bit of a D&O underwriting approach, there is some analogy to be made here. Underwriters and actuaries need to be less micro focused and adopt more of a macro lens to understand how best to price going forward.
With that said, I think we all know the industry will not move in this direction without experiencing sufficient pain first. Which means there will be sufficient pain in the years to come. If there’s one thing I am comfortable predicting, it’s that the industry will botch this transition before it gets it right.
There are so many challenges that face carriers but personally I wouldn’t put a return to 1970s-style inflation at the top of the list. While it’s true that today’s actuaries did not personally experience the inflation of the 1970s, it’s also true that the inflation of the 1970s had some unique characteristics, which are unlikely to recur. The sharp spike in oil prices was the direct result of the OPEC embargo, which was a reaction to the US support of Israel in the 1973 Yom Kippur war, a very specific situation. And we have since become energy independent and diversified our access to global energy resources. Also, at the time, there was a legislated very low cap on the interest rate on savings accounts (later repealed) that strongly discouraged savings and so reduced the funds that banks had to lend to businesses. Today we are net savers and businesses have, in theory, many sources of capital. Yes, the pandemic has created some supply chain challenges but … will they be long-lasting or will we read tomorrow of new US chip manufacturers, easing dependence on too-few sources? Even with sharply increasing salaries in many sectors, is this an inflationary long-term problem or an overdue and overall beneficial remedy to the systemic problems of income disparity?
I think the #1 risk faced by carriers is the unprecedented rate and scale of nat cat events attributable to climate change and, in this sense, the author’s thesis is right: actuaries can’t model it well so there are likely to be unexpected adverse consequences. Perhaps equally, severe shortages in specific areas will up-end actuarial models, like chip shortages/car shortages/repair shop & parts shortages, and have a similar economic impact as the oil embargo of the 1970s.
Bottom line, it’s tough to be a carrier. The drive for rate increases for a myriad of very valid reasons is seldom high enough or sustained enough to counter these risks.
Thanks Gretchen, I appreciate the thoughts. What I will fall back on is history doesn’t repeat, but it rhymes, right?
If you believe semis are the new oil, then Taiwan Semi is the new OPEC and it’s future is also clouded by geopolitical risk.
We are seeing energy itself now challenged by the mismanagement of the transition to renewables, particularly in Europe causing shortages.
And while interest rates aren’t formally capped, they are clearly restrained by Fed policy. The eventual end of QE/MMT could be as perilous as the end of Bretton Woods.
So there is a lot of rhyming potential out there. If I were a more talented writer, maybe I’d turn it into a song! 🙂
As for insurance specific issues, yes, the cat models are wrong, but that story has never changed! 🙂 We are only debating a new reason for them to be wrong!
I’d link back to my cyber article as another risk of trying to model the “unmodelable”, but I’m not seeking a repeat of the hate mail that flowed from pointing that out last time!
But yes, I think we are in agreement on the more general thesis that model inaccuracy risk is above normal and poses risk to the industry.