Is there supposed to be something inherently
not-fine with a conglomerate?
Some people think they have a tendency to be inefficient, as high-performing areas of the business waste money subsidising low-performing areas of the business. And low-performing areas are insulated from free market feedback, like going bankrupt or investors leaving.
Of course, some of this stuff is only obvious with hindsight. It was probably inefficient that Google Ads wasted money subsidising Plus and Wave; but probably efficient that Google Ads subsidises Gmail.
It's not about individual cases, but that having the businesses combined prevents investors from allocating capital between them, which is viewed as, in aggregate, less efficient as the allocation is a step removed from the market.
This assumes that investors are efficient at allocating capital, or that individual investors are more efficient than institutional investors.
As an obvious counterpoint: Birkshire Hathaway
A more local example, as an Australian, is Wesfarmers who have persistently maintained strong brands in hugely diverse fields. As a welder by trade I'm willing to bet that the people managing these two conglomerates are more efficient at allocating capital than I, at least.
> This assumes that investors are efficient at allocating capital
Yes, in aggregate, it does. Phrased another way, it assumes that the market is more efficient when different investments are unbundled and can be selected independently than when they are bundled.
Note that I am not necessarily endorsing this charge against conglomerates, just clarifying what I see to be its basis.
Of course, some of this stuff is only obvious with hindsight. It was probably inefficient that Google Ads wasted money subsidising Plus and Wave; but probably efficient that Google Ads subsidises Gmail.