It's actually the reverse. Buffett requires his insurance companies write profitable policies, so that interest earned on the float is gravy. When one suffers through an inevitable bad year (Gen Re a recent example) float interest is a greater margin of safety.

Most insurers write unprofitable policies and use float interest to cover those losses and make a profit. When they have a bad year, they often lose horrendous amounts and can be forced to raise more funding at unattractive rates just to stay in business.

And the value of float seems to be over-rated on this site. Float almost never can be invested in the stock market or other high yield aggressive investments. Because it represents policyholder's premium dollars it has to be invested very conservatively, usually in safe fixed income investments. The way Buffett uses it adds a nice bump to earnings, but it's far from a key to his success. He never owned any insurance companies when he had his highest return decades.

Float can be invested in pretty much anything. You're right that as an industry it tends to go into low risk fixed income investments but that's because it's safe (and easy for companies whose core strength is selling insurance) not because it's required.

Berkshire entered insurance in 1967; were their best years prior to that?

Buffett's best years were the Buffett Partnerships, from 1956 to 1969 (iirc), where his returns averaged 40% per year.

Berkshire has benefitted from insurance float for a long time, but it's a minor benefit. Float can't be invested in the stock market, it's extremely restricted in it's usage, and it's only been helpful because of the incredibly disciplined way Berkshire runs it's insurance companies. Proof is it never gave Berkshire Hathaway remotely the returns the Partnerships achieved.

That's not proof at all. 60 years ago markets were much less efficient than they are today, and Buffett Partnerships was incredibly small compared to the size Berkshire is now.

Buffett has said it ad nauseum- the anchor around his neck is the size. Beating the market with what $200 billion to invest? That's leagues different from beating it with $10 million.

That's an interesting argument because basically any modern Berkshire letter you read is going to say something along the lines of insurance being "the most important" part of BRK, an "all-important" component of their revenue, "the cornerstone of Berkshire", "our core business and the propellant of our growth since we purchased National Indemnity", "our main business", "a huge winner", "exceeds its book value by more than any other Berkshire business", &c.

(Can you tell I just plugged a bunch of random years plus "Berkshire letter" into Google? I didn't find a single one that didn't play up insurance; I'm not leaving any out.)

In 1994, he even laments exiting a position in GEICO in the 1950s as one of his major mistakes!

One of many fun parts in reading the BRK letters is seeing this "tweaking" of their underwriting through the 70's until eventually losses were rare.