Leaderboard

I was having a long conversation with an AI model about Fairfax's 10 year history of reserve redundancy / positive development of prior years reserves and how Fairfax has been very good at over-reserving up front, which has the effect of over-stating "float" and understating "shareholders equity" - since a fairly predictable chunk of "liabilities" is actually equity that just hasn't been released yet. This also has the effect of lowering the coveted leverage ratio (2.5-3x) we receive because equity is actually probably higher than stated. A similar lowering of the leverage ratio results from the under-marked assets we know are worth a few billion more than the accounting books say ("excess of fair value over carrying value"). Either way, Fairfax is working hard to shrink their equity through repurchases and maintain their leverage and they have a higher leverage ratio than Markel. Berkshire lost its leverage a while ago. Sad. The conversation moved into a discussion of the behavior of these prior years reserve releases (mostly in Q4s) in the 3-6 year period after a "hard market" and it became pretty clear that a whole lot of the past several years of hard market benefit has not actually been reflected yet on the books. A major counter-cyclical earnings buffer that makes those softer years a lot more enjoyable (another added bonus is growth slows, freeing up capital to be distributed down to the holdco). Because Fairfax grew so damn much during the hard market years, the annual contribution from reserve releases / positive development will likely be much higher than the $500m annual "typical" release and 2025's $751.5m. If you believe Fairfax has maintained the same over-reserved conservatism - and I don't know any reason to assume otherwise - then they already have a higher shareholders equity and lower liabilities than their accounting book indicates.