You claim an “economic” approach to testing and yet present no kind of mathematical model at all to support your supposition. You also misappropriate the word “Lean” in much the same way the “Lean Startup” community has, cherry picking a very specific lean idea in the pursuit of a foregone conclusion.
You wantonly misrepresent the TDD article: “There is some empirical evidence that suggests Test Driven Development (TDD) does not improve code quality” and present nothing but an article which even the title denotes is focussed on *design* quality.
I see nothing in this article other than I well constructed argument to obfuscate the lack of evidence for a pre-conceived position.
I don’t object to some of your conclusions, integration tests have a different value to unit tests, I object to the tortuous way you arrive at a conclusion you already held — you don’t like unit tests.