I believe the model is "100% pay for 80% of the hours and 100% of the output". This is usually accompanied by hand waving talk of "improving productivity" but I don't see how it will work realistically. If you were already spending 20% of your working hours scrolling Facebook (or HN) then you will do so on a four day week too (perhaps after an initial honeymoon period).

The alternative approach where employees just spread their 40h of work over 4 days instead of 5 is interesting too, but in a different way. Is the loss of productivity caused by being tired by hour 10 each day offset by having 3 full days to rest? Or even, perhaps those long stretches with nothing to think about but work are MORE productive? A 10h day could be tough on home life (kids/spouse/evening leisure), but if WFH that could just be offset against hours now not spent commuting. Everyone's situation is different. Will be fascinating to see the outcomes of this study. For starters, "define productivity" for a software business!