Finland Will End Its Experiment with Universal Basic Income | Hacker News Reader

Finland Will End Its Experiment with Universal Basic Income

time.com

33 pointsdbattaglia25 comments

Discussed earlier:

https://news.ycombinator.com/item?id=16874921

What good is an experiment applied to a small group of people and only for a limited time? The people that get it aren't going to adjust their long term behavior and the local economy certainly isn't going to change in response.

Well, we have no alternative. We can't afford an experiment that funds a ton of people and lasts for a long time. No matter how abundantly confident the people of HN may be that BI is the way to go, one of the fundamental lessons of science over the past couple of centuries is that your "abundant confidence" doesn't count for squat in the real world.

If BI can't prove itself out in a smaller environment, we on no account can afford to risk our entire economy on it.

That said, I'm not saying this was necessarily a viable test of the idea. On the other hand, if it really is a great idea that is going to save civilization, it really ought to have the characteristic that it is still a good idea that is an improvement on the status quo even when applied in the small, such that everybody wants to continue the experiment and expand it, not shut it down. The slow drip of failed BI experiments, even if they didn't "really test BI", is becoming non-trivial evidence against the idea. An idea that only works if we wake up one day and transition the entire national economy at once, or worse, even the entire global economy (which the combination of "open borders" and BI could require, if "open borders" includes handing out BI to everyone who crosses the border), is a non-starter. Even if in some abstract sense it would work, we can't ever attain enough confidence in the plan to make it a rational choice.

There are certainly things that have the curve that if done a bit make things worse, but make things better if done wholeheartedly. However, there's a lot more things that people think have that curve but actually make things worse the more they are done. It is impossible to tell the two apart from here, and the abundance of bad ideas should give one pause before making a huge commitment. The twentieth century is strewn with the wreckage of economies that made that mistake, and the twenty-first century hasn't stopped, either. We can really only consider interventions that work in the small for our changes in the large. (And always keeping an eye out for the interventions that worked in the small but failed in the large!)

Simply select 2000 18 year olds and give them BI for life. That isn't all that expensive even in worst case and would actually show if BI could work.

Pretty sure you can find 2000 18 year olds with a trust fund or part of a Native American Indian tribe that distributes gaming revenue.

"Coming from a rich family" or "part of a Native American tribe" is not a wide enough distribution of people to draw conclusions. Any failures / successes could be down to the environment and may not be able to be extrapolated to the rest of society at large.

On the other hand, if you're wondering if people who after the end of highschool, are able to finish college without having to get a minimum wage job to pay rent and for college, are doing better in life after college (and what measured amount better), the data already exists and there are far more than 2000 subjects.

It wouldn't at all show the effects of a large scale implementation. You can't measure the effects of supply constraints and rising prices with small scale tests.

If only it were simple. To actually do that, someone somewhere has to approve a budget of at least 2 billion dollars, that's how much money you're talking about. They have to both find the funding and have the political stamina to withstand the backlash on that kind of spending.

* my assumptions are: $25k income (which is just barely above the poverty line), for 40 years (which isn't long enough).

Of course there is an alternative. Fund 1/10 as many people for 10x as long. I feel you'd get a much better idea of what people would do despite the smaller sample size. The mentality of someone who is set for life vs one who has things taken care of for the next 5 years is completely different. Many of the latter are going to be thinking about what they can do to ensure employment after their "sabbatical".

>>No matter how abundantly confident the people of HN may be that BI is the way to go

I think people of HN are far from "abundantly confident" about BI. Heck, it's probably the other way around: every time the topic is discussed people come up with reason after reason as to why it's a terrible idea.

> …if it really is a great idea that is going to save civilization, it really ought to have the characteristic that it is still a good idea that is an improvement on the status quo even when applied in the small…

Not if we're currently locally optimized. Then it would seem worse before it seemed better.

I'm not particularly for or against the idea for my own sake. However, I do think it's going to be inevitable sooner or later. Whether it's a good fit for us right now, I can't determine.

In my opinion, experiments are needed not in order to determine whether it should be implemented or not, but how it should be done, and in what form. When the level of unemployment starts leading to riots, we'd better prepare with something other than weapons and harsh measures.

"Not if we're currently locally optimized. Then it would seem worse before it seemed better."

I think you don't quite have the point I'm trying to make, which is not surprising as it's the first time I'm trying to express it clearly. As I said, I'm sure there are interventions that are locally worse but globally better. Their existence is not the problem. But given the much, much higher prevalence of solutions which are locally worse, and then also globally worse, and the observable fact that many people believe these are good ideas anyhow (simply look across the political isle; everybody will disagree on what these ideas are but everybody should agree they exist), the challenge becomes one of identifying which are actually globally better even though they are locally worse.

And I submit that there is no solution to that. At least not for humans. Maybe for Vulcans, but not for humans. (Anyone who wants to disagree with that, please double-check that your argument doesn't amount to "It would work if everybody would just change their opinions to agree with me.", which I pretty much guarantee it will in the end.)

So our only rational alternative is to pursue solutions that are locally better, then to try to scale them up (again, keeping an eye out for if and when they turn negative, because they can), because there is no mechanism for discovering interventions which are locally worse but globally better. If the local experiments being done on BI do not have sufficient local success that all involved want to scale it up, we have no mechanism for rationally deciding "Well, it's not working so great on the small scale, but it'll totally work on the large scale." We could only irrationally leap to it, and we would rationally expect to suffer the usual consequences of such irrational leaps. (We are not scientifically or rationally entitled to simply assume BI will work in any sense.)

The other thing I'm trying to point out is that even if none of the interventions to date are "true BI", they are micro-scale BI, and if they aren't working there, it is rational to start considering that evidence against BI. BI also can not work if it must be run "just so". If it is only valid at a particular point in the policy space, it will fail. If it is going to work, it must be robust to perturbation. Policies that work in the real world have to be robust against exceedingly intelligent adversaries, and for all everyone may complain with varying degrees of justification about the current state of the world, the simple truth is that we do indeed have a lot to lose and there is nothing that stops us from losing it if we make wrong decisions. Collecting several samples even from the edge of BI and seeing them come back negative is not a good sign for BI ever being the rational choice.

What good is a social policy if you have to implement it at national scale for an unlimited time to see if it works as advertised?

That largely describes US Social Security.

Democracy worked out pretty well.

Democracy scales and in fact works better in small groups than large groups...

Democracy was tried in many different settings and scopes, over the course of many centuries, before gaining a real foothold. America was NOT a great experiment in a new approach to government - or if it was, that experiment was with a Constitutionally limited system.

It was implemented in ancient Greece as we're all taught. At the time of America's founding, there was already a democracy of sorts in place in the Netherlands. Along the way, many other institutions implemented it - for example, pirate ships were often democracies, with the captain chosen by popular acclamation.

Who said anything about the US?

The exciting stuff with negative tax/basic income is to see the systemic effects kick in. And yes, this will be very hard to test small-scale. Kind of how it would be very hard to test whether democracy works or not by giving one person a vote.

I'm only speaking to the value of the experiment. Maybe something can be learned by conducting it, but not about the viability of UBI. It was a waste of time and money in that regard. However, the people who have to deal with other people asking about UBI can now say "Our studies show UBI to be a waste of time and money."

I'm not convinced these big bang experiments, in the context of a wider economy not organised around the concept, are that useful.

For me one of this big advantages of BI is not enabling more people to 'drop out' of the productive economy and become artist such. It's actually the way in which it subsidises low income jobs. One of the big problems we have at the moment is low paying work being squeezed out of the economy through automation and increases in the minimum wage. A basic Income enables more people dropping out of the labour force to get back into it, because with a BI low paying jobs become a more viable way to earn a living. This gets people out of the poverty trap and also makes businesses dependent on low wage work more viable. It could be a real shot in the arm for the employment sectors that the most vulnerable people in our economy depend on.

On that basis, rather than pick a few thousand people to give a large BI to, why not start by providing a very modest BI to everybody. Fund it by equally modest tax increases, so it's revenue neutral. You could start with on the order of a hundred bucks (quid, whatever) a month. If a BI is going to have a beneficial effect, then even at this level it ought to do some measurable good.

If this was an experiment, what were the results? What was being measured?

According to the article, it will end in January 2019 (the government decided not to fund extending it), so results can be expected in a year or so.

It can be an experiment without being scientific or rigorous.

In this case, I believe the ultimate goal was to see how long they could keep it going.

EDIT: Why the downvotes? The word "experiment" doesn't automatically imply science. Why couldn't Finland have been experimenting with UBI the same way that individuals experiment with many things: to see what would happen?

Reply on news.ycombinator.com