My guess is because with working with text is more efficient. The cognitive load of finding a setting in the UI, moving your mouse to it, and selecting the value is far greater than just typing.
Perhaps, in part, because one allows for a greater expressiveness -- albeit with a little more planning and cognitive load.
Visual editors don't work so well with pages where parts are static & parts are dynamically generated. (Like, a css class may be different based on the data that it's showing). And at this point, pretty much every page in a web app has dynamic parts.
Visual Studio and many others have this out of the box but there's plenty of companies that do well selling exactly this like telerik.
I personally don't and been doing it the "hard way" since the beginning because it feels less confining. And I don't want to rely on an IDE to build something. Plus the code it generates was always kind of kludgy.