I actually do, at least with respect to hard tests, and let me explain why.
In the US, no one is “shunted.” Ultimately, the composition of classes for universities will be shaped by student choice. A high scoring student on such hard tests may, for whatever reason, choose to attend a school with lower average stats. And vice versa (assuming that enough higher scoring students choose not to attend a school in which high scorers appear to predominate).
As well, in response to these dynamics, some schools will find that their mission is best served by ignoring or limiting the influence of test scores, relying instead on more holistic measures. That’s fine, too. So long, of course, as they don’t discriminate on the basis of race. Race discrimination is and should be proscribed in the United States, no matter the incoherent dissembling of SCOTUS and the various interest groups championing such discrimination under the guise of virtue.
Nothing is static, and one of the first principles of “real” economics - not the standard bilge that is dispensed in universities - is that people act in response to available choice, so long as there is a choice and so long as they can act. Efficiency will be maximized when the number of choosers is large, and the suppliers of choices can be flexible in response.
The elimination of truly difficult tasks - like hard tests that can rank order thousands of kids even in the top 1% of ability - takes away valuable information, both to the kids as well as to the colleges. The absence of such information must necessarily lead to a decrease in the quality of decisions, so long of course as academic ability and intelligence are attributes that are still desired by at least some colleges. As information is reduced, choice progressively shifts to the suppliers of the choices (e.g., adcoms) rather the presumed beneficiaries (i.e., students). These suppliers will substitute their own whims and preferences. I’d rather see the opposite dynamic, especially as we see that certain prestigious colleges seem to systematically favor those attributes which are likely to be highly correlated with income and wealth.
The argument against difficult tests, although it has never been a particularly strong one, is that they favor the wealthy and sophisticated. Well, it’s actually pretty easy to address this deficiency by making the tasks more g-loaded, more sensitive to innate intelligence. The privileged today resist this idea furiously, primarily because they fear that their kids would not fare well.
The truly underprivileged today, on the other hand, have nothing to lose under this approach and will likely see jumps in representation at those “top” schools (the ones that ultimately shake out as the “top dogs” from this dynamic choice-response process) akin to those enjoyed by poor Jews when Harvard and Columbia got together to design the SAT about a century ago. That’s why the status quo will not allow it.