What ratio of PhD graduates in STEM fields ultimately end up as (tenured) professors?

It's a shiny info-gram, but I think there is a lot wrong with it:

• It transports the (common) misconception that the 53% leaving to industry after their PhD are sort of failed professors. At least in CS, and probably a lot of other STEM fields as well, many PhD students start with the full intention of leaving academia sometime. Hence, the better question the diagram should be asking is How many of those that want to become professors actually do?. Basically, I could come up with similar low percentages for every field. Let's say less than 1% of all working population of a city works in supermarkets. Does that mean that supermarket jobs are horribly hard to break into? No, because most people do not have the career goal of working in a supermarket in the first place.
• Similarly for the 17% non-university research. In CS, good industry labs (like the ones at IBM or Microsoft) are preferable options for many researchers, so they would not take a professorship even if offered.
• As already stated by commenters above, the title professor means different things in different places. In Austria, for instance, many high school teachers are officially "professors" (tenured even), despite not having a PhD or ever doing research. In Great Britain, very few people are professors (most are lecturers or senior lecturers). In Madrid I know an academic research lab (not affiliated with a university) that calls their staff members research professors. Job titles are almost never clear-cut.

That being said, I do think that there is a problem. If we assume a reasonably stable system (number of professors in a discipline stays more or less constant), then every tenured professor is in average allowed to see one of her/his students through to also become a tenured professor. Given that many tenured professors (at least here in Europe) maintain groups of 15+ PhD students at a time, I is pretty obvious even without digging into the data too much that the job market for professors is insanely competitive (which, incidentally, captures my personal experience hunting for tenure-track positions pretty well).

Just looked up some numbers for Germany:

• In 2013, 8700 PhDs were finished in maths/natural sciences.

• average age at finishing (all fields): 32.5 a

• average age at becoming professor (maths/natural sciences): 40.5 a

• predicted number of retiring professors in 2021 = in 8 years = when last year's fresh PhDs reach the average age of becoming professor
  (maths/natural sciences): ca. 190

190 : 8700 ≈ 1 : 46 or 2.2 %

Some of the tables show only overall numbers, and no details for maths. But I think that this result is influenced by the fact that the majority of chemistry and biology students go on for a PhD (though I guess that a non-negligible fraction leaves for industry [slightly] before finishing the PhD - which after all may not be that different from doing a PhD in order to get a better entrance position in industry).

Sources: Statistisches Bundesamt

• exam statistics
• university staff statistics

An article in Physics World has more information, and appears to be the source of the figure in question (I'm not entirely sure if it is the original source, as the article draws it data from elsewhere). The article is available to subscribers, and the full citation is:

• Harris, Margaret. "The academic pyramid." Physics World 25, no. 10 (2012): 54-57.

It appears to be presently mirrored here, and my information is drawn from the mirror. It answers some questions on data sources that were lost when the image got its own life without proper context. Data are for STEM fields, and are relevant for the United Kingdom. The figure caption reads:

Transition points in typical academic scientific careers following a PhD. Based on data from the Higher Education Funding Council for England, the Research Base Funders Forum and the Higher Education Statistics Agency’s annual “Destinations of leavers from higher education” survey.

Furthermore, the article states:

Statistics suggest that the vast majority of people who complete science PhDs will never obtain a permanent academic post. This is vividly illustrated in a diagram published in 2010 by the Royal Society as part of a report on the future of scientific careers in the UK (figure 1). Drawing on data from various UK sources, the diagram follows a “typical academic career” through a series of post- PhD transition points, when large numbers of people leave the university environment for careers in, say, government or industrial research. These data show that less than 0.5% of science PhD students will ever become full professors, while just 3.5% will obtain lower-ranking permanent positions as research staff at universities.

For physicists, that 3.5% figure is probably a little low. Slightly older data collected by the Institute of Physics and the US National Science Foundation suggest that the fraction of physics PhD students who obtain permanent academic jobs has historically hovered between 10 and 20%.

(...) But many more do want to stay in academia:

Indeed, according to an August 2012 survey carried out by the American Institute of Physics (AIP), nearly half (46%) of new physics PhD stu- dents at US institutions want to work in a university. The next most popular career plan among those surveyed, attracting 18% of responses, was “unsure”.

For more information, the article points to the UK group Vitae, UK science advocacy group Science is Vital, and the US NSF Statistics page.

So, for physics, it appears between ¼–½ of PhD students who want to get permanent academic positions, ultimately succeed in doing so. That's a quite different figure than 0.5% (but still problematic, as the article discusses in some detail).