Shane's conclusions regarding public policy run counter to most of what I have read. He says:

"We need to reduce the incentives for marginal entrepreneurs to start businesses by reducing the...tax benefits that encourage more and more people to start businesses. Because the average existing new firm is more productive than the average new firm, we would be better off economically if we eliminated policies that encourage people to start businesses instead of taking jobs working for others."

One of his premises from earlier in the book is that most startups in which barriers to entry are low are colossal wastes of time. This may be valid but it does not follow that GDP would improve if we made companies even harder to start.

Notes from the book: http://dl.dropbox.com/u/17455578/hn/illusions-notes.html