There are several reasons for this point of view:
- when looking for a survey/review paper in a journal or talk at a conference, the first thing I do – is looking at the name of the researcher, how good (and how many) are his works in this field (and subfield of the overview talk).
- to give a good survey talk, you both have to have a "big-picture" vision, as well as exposure to various details in many methods. It is hard to achieve both at the early stages of your career. And, most importantly, your potential audience has a tendency to think that it is unlikely.
- giving review talk is, in my opinion, harder than a novel research one. The structure is different, and you cannot focus on the novel idea that is your baby and you are presenting. I would say, you need better presentation and slides' organization skills to prepare a similar quality survey talk.
- presenting survey talks and publishing review papers is not perceived as good for your academic career. While it counts as a publication (and is no doubt useful for the research community), it is no equal in terms of the impact on your CV. Surveys can rarely define somebody's career and they play less of a role during application for a tenure-track position.
Is it fair? No. Can a young researcher give an excellent, inspiring, and even novel overview talk? Sure. Can that boost your research reputation? Sure. Still, unlikely in general.
I can totally see why the preparation of the survey paper itself is very useful:
- by preparing a review paper you learn the subject matter, past and recent progress in the best possible way
- you create a quality publication for others to enjoy
- you learn the skill of writing a survey paper itself (which is different from a regular paper in a journal)
However, in my opinion, the benefits of actually presenting such a paper at a conference are limited for an early-career researcher.