Opinion: What to know about social inflation, insurance

Puneet Prakash

The phrase social inflation has been used in the insurance industry since the 1970s, and Oracle of Omaha Warren Buffett was among those who used it. The phrase raises its head when the insurance industry witnesses abnormal losses followed by a contraction in insurance supply and an incidental increase in premium rates, especially in the long-tail lines of liability insurance.

While the phrase lay dormant between 2000 and 2017, it took center stage by 2020 because an increase in liability losses incurred. But it never gained traction in the popular press as it lacked a clear definition. 

In the ’70s, the usage pertained to large jury verdicts. Defining it as narrowly today, in terms of losses due to “nuclear” verdicts with $10 million or more in payouts, invariably leads to interesting arguments between legal and insurance practitioners with each side blaming the other for losses.

Insurance companies point to litigation funds, jury and jurisdiction shopping and aggressive class action suits by lawyers, while plaintiffs’ lawyers blame incomplete actuarial models, marketing and claims practices of insurance companies that are unfavorable to consumers, thus forcing them to turn to the judicial system for indemnification.

In its latest avatar, social inflation can be defined as the increase in insurance losses caused by legislative, judicial, social and economic, and technical developments.

Even with a broad definition, measurement of losses attributable to social inflation is nearly impossible. Some have tried to measure it by the excess of what can be explained due to price inflation, while still others consider excess of “normal” losses. In essence, losses due to social inflation are measured as losses beyond some expected benchmark. While short-run deviations from this benchmark will occur, the long-run losses are supposed to equal the benchmark. Thus, the choice of benchmark is crucial. In practice, this equals the expected future losses plus some cushion for the short-term deviations.

So, why did the phrase social inflation appear at high frequency after 2017? Data by industry researchers show that incurred losses in all liability lines grew between 3.2% and 17.4% on a year-over-year basis between 2013-14 and 2017-18. Meanwhile, the rate of inflation over this period was 1.5%. Using price inflation as a benchmark would appear to mean social inflation was the cause.

However, insurance premiums are based on losses expected in the future, so benchmarking against inflation inherently assumes that future losses are perfectly foreseeable, which is not the case. This line of argument would suggest that social inflation may be a real cause for an increase in insurance premiums. But why then did the popular press not buy into this phrase?

The reason is that it’s extremely challenging to disentangle the effects of social inflation from the economic phenomenon of insurance underwriting cycles. What the insurance industry attributes to social inflation could just be a result of a plethora of industry practices that accompany the cycle – for example, underpricing and laxity in underwriting standards to gain market share.

Following large losses, hard markets in underwriting cycles are characterized by increases in premiums with lower coverage limits, leading to higher insurance rates. The opposite is true for soft markets. But that is only the first-order effect.

The second-order effect is on capitalization. Firms that raise rates higher and faster relative to their competitors in a hard-market environment are the ones that were undercapitalized to begin with. And empirically, firms prefer raising rates than going out of business in a hard market. To the extent that social inflation is totally exogenous to insurance underwriting cycles, one expects to see mass exits of firms of all sizes from industry. However, that has not been observed despite shocks due to the pandemic. So, while the insurance industry professionals can continue talks of social inflation as an emerging risk that is affecting their losses, popular business press is likely to glaze over it and consider it yet another insurance cycle.

Will the industry ever be able to disentangle the effects of social inflation from other causes of the underwriting cycle? The current advances in big data with artificial intelligence, machine learning and large language models are being used to predict expected losses with increased precision and may just help in this task. Only then can the industry begin to try convincing the world outside that social inflation is real.

Puneet Prakash is a professor and Baker chair of risk management and insurance at Missouri State University. He can be reached at puneetprakash@missouristate.edu.