There is, of course, a third argument: In the presence of significant specificities, do not go into the business at all. To justify the overhead of vertically integrating, the business must be very promising, and yet by nature the uncertainty makes it