Social Inflation, is this real?

Yes, but it not constant over time. The predictions you made on one particular historical data set are not guaranteed to be consistent with the data in the near future.

Furthermore, in some areas/lines of business, the effect of changes in inflation/claims severity are EXTREMELY leveraged, like umbrella and reinsurance. In those areas, the cost missing the claim severity estimate could easily cost a company tens of millions of dollars, and perhaps more. I mean that literally and without hyperbole.

I’m not bloviating. I am trying to keep my job.