At the risk of revealing my ignorance, wouldn't it be feasible to set up an intelligently designed array of subsidiary domains for the purpose of acquiring and managing sets of smaller companies?
...
This is such a simple answer that … it was the first thing I thought of too. Clearly there’s something wrong with the idea, though what that obvious defect is escapes me.

There are two problems that I can spot.

First, this is now a very crowded field. There are many, many BDCs and PE folks trying to buy good cash flow generating firms of small/medium/biggish size. So much so that the average quality of listed small firms has fallen through the floor...there is no longer a pipeline of good ones. In short, the goods on offer aren't great and the prices are high.

And second, who would run such a thing? I'm sure a younger Mr Buffett could do a smashing job even in this overpriced market, but he is is short supply.

The better choice might arise if there is an opportunity, perhaps during a market disruption, to buy a business already running such a portfolio. The operating-subs division at White Mountains, something like that? (Not that I expect them to have financial difficulties, just an example of a firm that has added a few operating subs). Maybe one of the oxymoronic public private equity firms that runs into unsurprising trouble but has some decent actual firms inside it.

Jim