It makes sense in a strange way. For a disaster or war, you put as few people in harm's way as possible, and overdub the reports in other languages for non-English markets in the studio. Viewers need facts and often the images speak for themselves, and the nature of the situation justifies endless replay of the same bits of footage.

For showbiz/celebratory type stories, almost all the reporting (in multiple languages) is talking to camera because viewers who care about that stuff want the experience of being there, so they enjoy it vicariously via the reporter. Also, there needs to be a much larger flow of material because people who want to be entertained get bored more easily.

There's a huge audience for this sort of thing. I was reading an economics paper a few months ago observing that the wedding of Charles and Diana generated the equivalent of 8 or 9 billion dollars in today's money, over the course of 20 years. This upcoming one is expected to add something like a billion to the UK economy just this year.