It was not a huge difference. They were not less productive, but the measured increase in worked hours was something like 6% if I don't remember wrong. They could take short gigs without losing unemployment benefits for the time and have to deal with the hassle with going off and on the benefits for just a few days of workt. But too small and limited study to really say anything about it.

The important gain would be to give people breathingroom to actually find a job most relevant to their education and skills instead of just jumping on the first available opening to be able to pay the rent. Or even spending some time improving your skills.

In the US I think just introducing universal healthcare would accomplish part of the same thing.

Since the target audience was rhe unemployed, changing up their type of income isn't going to give them a job, so prpductivity is difficult to measure.

The entire point of UBI is to secure everyones basic human needs, not get them a job (unlike unemployment benefits).

If we were to test UBI also on employed people, and observed their behavior, we might avtually gain useful insight into mechanica with UBI. Perhaps we will see more 6-hour workdays with more people employed?

If you already have your basic human needs met then it is just a shortcut to provide the money only to those that don't have that one checked off yet. Otherwise you'd just end up pumping money around from those that don't need it to those that don't need it.

Employment, the only measure of productivity that is considered in this context, was (maybe) modestly increased or otherwise not affected.

Given that it significantly improved mental health, it seems reasonable to expect that this UBI model would lead to some savings in public health expenditure. Arguably there should also be some savings compared to the administration of the current unemployment benefits that require the person to provide evidence of activity and other such things.

How well it compares in practice to the cash cost nobody knows as the experiments are all too short to really tell.

How would measure that? Do you only count paid labor? Is all labor productive?