True, measuring productivity is notoriously hard, especially in areas like software engineering. It's largely an unsolved problem. So we have to rely on intuition, tradition and gut feelings, unfortunately. All of those suggest "let's work less by 20%" is probably not the most obvious way of improving productivity. I mean, of course, there's a place where working more does not yield more output. But 4 days a week doesn't seem to be that point - if fact, it seems pretty luxurious point to be in, and, again, intuitively, it looks like working a bit more would still produce marginal increases. It very well may be that the intuition is wrong - but to claim it's the case one need some serious proof. "I feel happy not working" doesn't really cut it - I'd probably feel happy working one day a week, for the same money, but my productivity would hardly be the same, so I don't see how I could sell such arrangement to my employer.