Since human capital (Adam Smith's labor component) is becoming more and more important as economies develop, sorting talent has followed suit. It becomes a concern of economists to ask if the people with the potential to become the best doctors and pilots and software engineers are in fact becoming those things, or if there are barriers (passive and active) preventing this. No doubt we are nowhere near 100% in any field; education is still uneven, there are still active barriers like gender- and ethnicity-based discrimination in most parts of the world; and there are passive barriers, like crude talent-detection tools.
I'm just finishing my third year in medical school, and I've noticed something interesting about how people choose specialties (or how specialties choose them). A very brief description of how medical education works in the U.S. is in order: first, you go to college and get an undergraduate degree in something, usually some science. Then you apply to medical school, which lasts four years. During this time you get a general medical education with experience across most fields and specialties, regardless of whether you're pretty sure you know what you want to go into. (The surgeons have to do psychiatry and vice versa.) Then when you graduate medical school after those four years, you are a physician, but you haven't been trained in any one area. Toward the end of those four years of medical school, you choose - do you want to go into family practice, neurology, emergency medicine, etc. and you apply for a residency, where for 3-7 more years you are trained in that area. At the end of that, you are finally a family physician, or psychiatrist, or surgeon, or whatever.
Whether and where you get into medical school is based on your GPA as an undergrad and your MCAT, which is a general science and writing test. And it's my impression that this process does its job fairly well: medical students seem highly motivated, highly generally intelligent people capable of doing well in medicine. But what I find disillusioning is that there is not a really strong talent detection tool to select for aptitude at the next step, when people are choosing their specialties. So how is it determined what specialty you go into? For the most part, you choose what you liked the most during your third year of medical school, and except for the most competitive specialties, most American medical graduates will get into a program in their chosen field, somewhere in the country. What differs is how competitive of a program you get into, which does affect where you go from there. And the inputs there are the disillusioning ones. It's basically, in this order: First, how well you score on the first part of your boards and how competitive is the area you're going into. Second: did the people you worked under in that specialty during medical school like you? (Yes, it's pretty much that subjective; there are grades but they only vaguely reflect effort and competence.) A distant third: did you not screw up the specific test they gave you at the end of the relevant rotation in medical school? Notice how little effort there is to really discriminate the specific talent of the person in that particular specialty among their peers, and how under-emphasized it is in the grand scheme.
I'll address #2 first. Microsoft commissioned a study where they determined how long it took people to form their impressions of a professional's competence, and found that they often do it before the professional opens his or her mouth, based (apparently) entirely on how much the person matches preconceptions of the appearance and behavior of a someone in that profession. So imagine my chagrin when I overhear, repeatedly, students being told by their superiors that they (the superiors) thought the student was going into specialty X, because "you just look like an X". (Seriously.) No doubt personality match and culture in each specialty make a huge difference to how engaged and effective someone will be during their careers, but this seems to be missing something important. Older students are often valued in psychiatry, and fortunately my "life experience" will benefit me in the specialty that most appeals to me. But if I had innate talent as a surgeon and chose that path, I would have had a hard road, because a guy my age with a fully formed ego not bound up in medical status hierarchy doesn't fit the idea of a young nose-to-the-grindstone surgeon in training. I don't look like a surgeon, regardless how much I would've wanted to be one. (Fortunately for all involved I was never in any danger of becoming a surgeon.)
As for #1: certain specialties are much harder to get into than others (ortho surgery and dermatology are extremely competitive for example). So in a way, your board scores do determine which specialties you can go into, but only by how good they are. (It's not like programs really break the score down by subsection, they just look at the main number.) So if you do really well, you have a shot at anything you want. If not, well, you're not going to be a dermatologist. And here's the interesting thing: 20 years ago, you almost couldn't pay people to apply to derm residencies. But then the reimbursement structure of medicine changed quickly, plus medical students gradually got wise to this and realized that in derm, here's a specialty that had excellent hours and great pay, and to hell with the status hierarchy. What this means is that (no offense to the great dermatologists I've worked with) based on their scores, your average dermatology resident today is more competitive globally than a derm resident 20 years ago. Is this appropriate? That is, have there been so many advances in dermatology in 20 years that derm residents have had to get smarter to keep up, or is it something completely separate from the field itself - and therefore, the talent of the people going into it? This is all to say that stratifying applicants based on their general board score is not finding the best dermatologist or OBGYN. It's sorting people based on the general aptitude, and then there will be a tendency for the top scorers to go into whatever is perceived as high-status and/or well-reimbursed at the moment.
Two asides. First, letting people choose what they like is probably the best part of the whole process, although this doesn't tell you how many people are choosing on status and money in the specialty rather than talent. Also, obviously there are many medical skills you can't evaluate with a traditional test, so to really find out who the surgeons are (for example) vs. the medicine docs or the neurologists, we would need simulations. Yes, this is difficult and imposes more time and expense on the medical education process, but I think it's worth it so that institutions (and medical students) know that the best-suited people really are going into each specialty, and anyway we already do have a patient simulation test for step 2 of our board certification.
My point: of course the reality is these sorts of inefficient sorting mechanisms exist many places in life. But in human-capital-intensive fields like medicine with socially valuable outputs, you would hope that the search mechanisms would be more robust. I suspect this is partly an artifact of antiquated medical education conventions, and my hope is that it will improve.