I don't think BV is losing it's relevance.

It still seems to be pretty good. It's a somewhat flawed back-of-the-envelope method, but it always was.

The bigger concern:
There have long been reasons to suspect that the fair P/B ratio might rise slowly over time.
But if anything, I think the fair P/B has fallen somewhat because the operating earnings have been quite poor for quite a while now - distressingly little progress in several years. For example, whatever the fair P/B is for BHE, it's lower than it was: the valuation multiple on those assets isn't what it used to be as they simply can't earn as much as they used to. Earnings at the railroad have been on a bumpy slight downtrend for a few years, a fifth lower in real terms than the peak rolling-four-quarter stretch about 4 years ago. Maybe it's a transient slow stretch, but it still isn't happy.

And of course the notion of applying a valuation multiple to the cash pile is based on the notion that any "higher than usual" portion is going to be allocated profitably within a reasonable amount of time. At some point even the most die-hard bulls have to question that thinking to some degree.

Investments per share have been rising very nicely lately, which has helped to offset the lack of progress on operating earnings, but any two-column type valuation metric is lagging any book-multiple type metric. Mine have diverged, anyway. My estimate of real "steady things" earnings from operating groups has risen only inflation+0.4%/year in the last 3 years, and that's entirely from the mild tailwind from a few buybacks in the first year. And times have been pretty good.

By "steady things" I mean the sum of after-tax real earnings per share from rails, utilities, manufacturing/service/retail, and a cyclically adjusted estimate of underwriting profit.

Jim