The core premise and compelling logic of BI comes from the fact that it is simple, not-gameable, and, most importantly, doesn't have huge disincentive effects at the margin as benefits phase out.
PER sounds like it runs headlong into all of those issues. By the standard you're using to critique BI using PER, I could just as well point to PER to critique markets as a whole.
There was obligation to work, unless you were sick. You could choose your trade, and if you were lazy/stupid/handicapped you were offered job that weren't demanding. Salaries were centrally planed, as was prices. The end effect was - less than 1% unemployement, huge drop in efficiency, no creativity boost to speak of.
I wonder if BI experiments suffer from observer bias - if the experiment was going bad it would be stopped, or at least people participating may think it would be. It's usually small community so they know who isn't working, and they can influence him.
If you were to introduce BI on huge scale, in a way that people take it for granted, I think the results would be much worse (and much more realistic).
People react to incentives. In the Soviet Union, "they pretended to pay us, and we pretended to work": although workers certainly received money in exchange for labor, increasing the amount of money you had did not typically increase the goods and services available to you, as the price system was so messed up that real prices were measured in waiting times and political favors.
A basic income avoids that trap because it leverages the distributed computation of the market to set prices. People will still be at least as entrepreneurial as they are now, because there's the opportunity for huge, outsized profits for both the entrepreneur and the provider of capital.
Also note that most salary workers nowadays make far in excess (typically 50k+) of what the minimum income would provide (on the order of 10k). Most of us here could work for a year, survive on what the minimum income would provide, and then have enough savings left over for nearly another decade at minimum income levels. But we don't do that, because we want more, and that wouldn't change with a minimum income.
I try to follow BI, and although the BI experiments I've heard of certainly have had cons to go along with the pros (spiraling costs in Manitoba, for instance), "stagnation" has never been one. I'd love to see that study.
[1] They eventually moved to Brazil instead.
The abstract suggests that it is looking at means tested and conditional benefits, including programs to "guarantee a minimum income": that should not be confused with a basic income, despite the similarity in terminology.
In the literature when people refer to a guaranteed minimum income they're referring to the government supplementing a person's income up to some predefined level. This makes it significantly cheaper than a basic income, because people who "don't need it" aren't receiving the income; but a really nasty side effect of that, as with nearly all welfare programs, is that it creates incredibly steep effective marginal tax rates. The simplest design--looking at someone's income, and filling it up until it hits the desired minimum--in fact leads to a 100% marginal tax rate, because a dollar earned from work is a dollar lost in benefits.
This means there's absolutely no incentive to work before you've made the minimum; contrast it to the unconditional basic income, where even someone who's only capable of producing half the value of the basic income is fully incentivized to do that work. So you avoid the deadly (literally!) welfare trap where the government actively punishes you for finding work.
I recognize the difference between BI and GMI as strategies but just so you know they are often both talked about in the same papers as they are both approaches to putting capital to work effectively amongst the poorest.
The challenge is to create productive work for these people so that they can earn an income. Some of the Chinese 'make work' projects have been interesting in that regard but they are not a final solution.
What might be a more durable system would be to create a manufacturing plant that these folks could work in. It doesn't have to be complex, could be as simple as making cinder block for building construction. It has to be run as a business in the sense that standards are set and checked and people are paid based on their production. You have to be rigidly scrupulous about keeping corruption out of it, this is a lot harder than you estimate. Ideally the production is consumed in other projects. (like building buildings or railroads or other factories).