The signaling model of education is surely neither 100% false nor 100% true, therefore it seems likely that your program would serve its function (getting certain kinds of people employed) better than no college and worse than completing college.

(Maybe this belongs in a separate comment, but I would add the further remark that your program works only if prospective employers are comfortable with and self-aware about the fact that they consider education to serve a purely signaling function. Since I think very few employers are really comfortable openly saying that they think college is not by and large a value-added proposition, I don't think your plan would work on a very large scale.)