The Dutch benefits scandal: a cautionary tale for algorithmic enforcement
eulawenforcement.com
eulawenforcement.com
Long story short, this article contains some grains of truth but the reality is much simpler: the Dutch tax office systematically engaged in discrimination, persisted in doing so when this was pointed out to them and made extraordinary efforts to hide their failings rather than to fix it.
The most ugly bit: the government that was in charge during all this fell due to the scandal and then was promptly re-elected claiming that they have full mandate to pursue their various un-ethical goals.
Families broke apart over this, children (1100!) got placed out of their parent's homes (because having your parents labelled fraudsters gives extraordinary powers to the government to really mess up your life as if it wasn't enough yet). A number of people committed suicide, one of which was directly traced back to the scandal.
If not for a single lawyer (González Pérez) persisting this whole thing would have never been brought to light. More than the algorithm I would blame the people who willingly implemented all of this. FWIW this is still unresolved, very few people (if any) have had any kind of compensation and the government is dragging its heels every step of the way.
All of that was made worse by the reluctance to admit there was a structural problem at all for years. That was the second-biggest problem.
Everything else are also problems, but this is really the root of the problem. It's worth pointing out that people with immigration backgrounds were far from the only victims, unlike your comment suggests (possession of dual nationality was the biggest "risk marker" they used AFAIK, not name, and it's not entirely clear just how biased the results were because people with migration backgrounds are also overrepresented as receivers of benefits).
In short, it would still have been an enormous scandal even without any ethnic profiling. That is not to say that ethnic profiling wasn't a problem, just the lesser one out of several.
> FWIW this is still unresolved, very few people have had any kind of compensation and the government is dragging its heels every step of the way.
This is genuinely a difficult problem. 1,379 people are involved full-time just for the compensation[1] There's even been criticism it's too easy to get compensation, even if you weren't a victim (see e.g. [2]).
Should it be faster? Yes. But it's hard to see how that would be realistically possible. I certainly don't think that the government has been "dragging its heels every step of the way" for the last 2 years.
[1]: https://www.rijksoverheid.nl/binaries/rijksoverheid/document... page 43
[1]: https://decorrespondent.nl/13592/na-de-toeslagenaffaire-is-n...
The criticism that it is too easy to get compensation is ridiculous, the compensation has two components, one to level any financial damage done in the past, the other because people were simply the victim of profiling, which the government isn't supposed to engage in to begin with.
Ethnic profiling was the problem, the root cause of the problem and continues to be the problem (the whole thing was triggered by the so called 'Bulgarenfraude'), there are also victims without an immigration background but those are by far in the minority and tend to be better represented with regards to the redress process than those that have other non-Dutch backgrounds.
Without the ethnic profiling we'd be looking at a small fraction of the number of people affected.
https://nos.nl/artikel/2434452-meeste-toeslagenouders-hebben...
Gives a really good overview in a small and simple table, the discrepancy between the chances of being marked as fraudulent based on ethnic background is off the scale and can not simply be waved away as not being the factor here, 70% of those affected has a migrant background and yet they only make up roughly 1/4th of Dutch society. It would take an extraordinary set of facts to explain this away and no satisfactory explanation has been given to date.
I don't doubt that it is unfairly biased, given that nationality was one of several "risk factors", but it's not entirely clear how unfairly biased it is. That table from the NOS is certainly not the full picture.
It's hard to tell what would have happened without any ethnic profiling. Maybe the group would be smaller, maybe not. Even if we accept the 70% figure at face value (which we shouldn't) and subtract all of those (which we also shouldn't) that still leaves us with ~7,800 ethnic Dutch victims (instead of ~26,000). Realistically, I think "at least 10,000" would be a reasonable conservative estimate. It's less than 26,000, but still a lot and very significant.
The way I see it, it boils down to:
1. There is no ethnic profiling; people still get trapped in a bureaucratic nightmare (possibly fewer than with ethnic profiling).
2. There is ethnic profiling; certain people get investigated more frequently, but people are treated fairly.
Both situations are highly undesirable, but situation 2 seems significantly better.
I agree that ethnic profiling was one of the sins here, but everything considered it wasn't the sin that led to the greatest amount of problems for the victims. I think it's a mistake to focus mainly on this aspect.
> Even now there are people here claiming that 'the bulk of those 30,000 must have done something wrong otherwise why did they end up in the database' and other such gems.
I can't recall ever seeing such comments, although I don't doubt they exist it seems a fringe opinion.
[1]: https://www.volkskrant.nl/nieuws-achtergrond/rammelende-data...
For starters the numbers would be much, much lower and the balance would reflect dutch society instead of this weird make-up.
> Even if we accept the 70% figure at face value (which we shouldn't) and subtract all of those (which we also shouldn't)
Well, you should: this was fairly extensively reported and if it were wrong I'm pretty sure it would have been seized upon immediately.
> that still leaves us with ~7,800 ethnic Dutch victims (instead of ~26,000).
There are no figures about the actual fraud but there is a good argument to be made that given the lopsided distribution the actual fraud percentages amongst the various people represented in the register does not reflect reality. If proven this would really put the lie to this system: the profiling not only accused a lot of people wrongly, and caused them immense grief and damage it also caused a lot of manpower to be wasted that should have gone to other cases of actual fraud.
The Dutch AP (Autoriteit Persoonsgegevens) is pretty clear in their findings:
https://www.autoriteitpersoonsgegevens.nl/sites/default/file...
> I can't recall ever seeing such comments, although I don't doubt they exist it seems a fringe opinion.
Have a look at the comment threads in Dutch media, underneath just about every article you will find these. Definitely not a fringe, unless you count double digits as fringe.
You seem to have completely ignored the parent's comment about how those immigrants are poorer and thus more likely to receive benefits. This is bad for being convincing.
The Volkskrant article I posted in my previous comment specifically addresses the figures in that report. The figure are not wrong, it's just more complex when you factor in everything.
> Have a look at the comment threads in Dutch media, underneath just about every article you will find these. Definitely not a fringe, unless you count double digits as fringe.
I consider those "fringe". I stopped engaging a long time ago as it's boring and pointless; I wrote about this some years ago [1] (the local newspaper being Eindhovens Dagblag):
> I stopped engaging on my local newspaper’s website years ago because the same small super-active group kept ranting about “the Muslims” on what felt like every fucking story, even when it had nothing to do immigration or religion. I engaged in good faith at first, but it got very tiresome very fast. For these people, everything seems to be about the Muslims, and there seems little interest to engage in constructive discussion. So at some point I just gave up. I’m almost certainly not the only one.
I think you need to be very careful going on random internet comments to see if something is fringe or wide-spread.
With a population of ~18 million people even just 0.1% is still ~18,000 people, and if this small group is very active and motivated they can appear a lot larger than they are – even more so if their nonsense drives out people who get tired of them.
FYI: This paranoid conspiratorial thinking is making it really hard to take the rest of what you say seriously. If you want to be convincing you should probably leave it out and focus on things you can prove.
You can't just compare the victims against 'Dutch society'. Not all of Dutch society has kids, and uses daycare, and applies for benefits etc. To make a sensible comparison to find out the amount of ethnic profiling that was used, you should compare the ethnicity of the victim group against the total group of Dutch society that:
- used daycare, and in particular the 'gastouderopvang' that were under suspicion by the BD, and
- applied for benefits (kinderopvangtoeslag)
Those who are fortunate enough not to rely on benefits may not realize it, but if you look at the amount of bureaucracy surrounding benefits, it becomes clear that this problem permeates the entire system.
If you rely on benefits, you are treated with suspicion by default (by design).
The same with 'kinderbijslag', 'zorgtoeslag' and lots of other of these social catchnet essentials. They're made into a massive meatgrinder and one little mistake and you'll be paying back the original assistance + some contrived fine, possibly much higher than the original amount you received.
My point: A lot of these benefits help children (directly or indirectly), who, in my humble opinion, deserve extra-special treatment, as they do not control their financial circumstances.
It's of course more difficult for the newly immigrated and the undocumented, but that's the case in practically every country in the world and the Netherlands is probably not any worse than other developed countries in that regard.
For one example: a child living abroad with one parent whilst the other half is living in the Netherlands has theoretically as much right to that support as one living in NL, but in practice the hoops they have to jump through + the risk of having to pay back the money increased with some arbitrary (but substantial) fine is so large that many people in that situation end up foregoing the aid.
It all works well when you are on the happy path, but if you are even a little bit outside of the 70% case ('normal') then the system will work against you in all kinds of unpredictable ways. And some of those ways would have you labeled a fraudster, even if you do your utmost to play by the rules. And finally: the government itself doesn't know what the rules really are so if you call one of their hotlines you may well get faulty advice that will eventually cause you to be labeled a fraudster.
There is quite some administration around applying for welfare ('bijstand') and also around daycare subsidy (mostly because it requires correct hours registration, and parents tend to change the hours quite a bit due to holidays, sickness, etc). But the administration around specifically 'kinderbijslag' (child benefits), 'zorgtoeslag' (health care subsidy) and 'huurtoeslag' (rent subsidy) are minimal to almost non-existent.
I disagree. The ‘toeslagen’ system is specifically set up such that if you start earning more partway through the year, you might have to pay months of benefits payments back (since those payments were really just a ‘voorschot’, and your actual right to them is based on your annual salary, which changes if you get a better job).
If you forget to update your salary (which might happen to ordinary people for any number of reasons), you might additionally be slapped with a fine.
I understand that on HN, we are self-selecting for people who understand the rules and how the system works. In real life, though, many people (especially those with a low income) are not well-versed in bureaucratic rules.
Consider, for example, that 1 in 9 (!) Dutch are functionally illiterate (laaggeletterd) [1]. The system is just too complex for them, and it’s no wonder many choose to forgo toeslagen when they do have a right to them.
[1] https://www.rijksoverheid.nl/onderwerpen/laaggeletterdheid
That said, for health care and rent it's not that hard, or at least it wasn't when I was eligible for it >10 years ago. For most regular people with a labour contract the tax service already has all your details and you don't really need to do anything. They also do try to use simple and plain language when possible (I remember that specifically as I felt it was too simple at times).
Child care seems different, which is why the entire scandal is mostly about that; I have no experience with that. And maybe the rent/health benefits changed in the meanwhile. The biggest caveat is that you may have to return the money if you earned too much as you mentioned (which also happened to me because I worked slightly too much overtime; I would have been better off not getting paid for that :-/).
If you rely on benefits, you are treated with suspicion by default (by design).
Exactly, and that is why I'll always bring up the only example of a benefit system that doesn't suffer all those beaurocratic nightmares. Polish child support (so called 500+) is a benefit for everyone that has a child(or children). It is 500pln (a it over 100eur) per child per month. It has been genius to give it to everyone. Minimal beaurocraticy was required as the state already has a record of children and parents. Fraud is minimal and it eradicated child poverty in few months.
There are still a lot of people hurt very deeply that "rich people get it too". Most people don't care, and it's goals are accomplished regardless. The cost is comparable with similar income-weighted benefits in other countries.
Edit:To clarify. Isn't current Polish gov right wing? Such things only happen in PL :-) That a right wing gov. brings a new benefit and the left is up in arms (because it has no income applicability). Now 8 years later the left has kind of accepted it was a good idea, but in the beginning they were very much against it.
This is why left/right wing don't always make sense in today's political climate. A radar/pizza chart (with the different categories - economic, social, ecological, etc.) is more appropriate than a slider but that doesn't provide for easy labels. A party can be right-wing, but pro-green stuff (renewables, decarbonisation, etc.) which isn't usually associated with right-wing politics. And of course populism makes it so that any party can implement policies which are widely accepted (such as giving money for children) regardless of (nominal) party colours.
This is the difference between a social insurance system and benefits. The former is much easier to reason about, and therefore can have simpler rules.
Nobody complains that "rich people gets pensions too" because of course they do. They pay taxes like everyone else.
This is well proven, and perhaps most well known, in the Nordic countries. When people treat certain measures as rights and not handouts, couples can plan their lives accordingly and manage both qualified educations and children.
This is also true for the 'kinderbijslag' in the Netherlands, the parents post about this is simply untrue. The vast majority of parents get it automatically in their bank account without any administration beyond establishing that you're a parent with valid resident permit and children and your bank account. There is of course extra work if you don't reside in the country or your child was born elsewhere / recently immigrated, but I can't imagine that that's much different in Poland.
> By far the bigger problem is that once you got labelled as
> "fraudulent" it was nigh-impossible to prove your innocence,
> no matter what kind of proof you provided
I won't go into all the boring details, but it was basically entirely my own fuck-up. I'd gotten some administrative detail wrong, and was simultaneously in the process of forming a company during that calendar year.
I signed up for a corporate hosting/help service that unbeknownst to me (yeah, I should have read the fine print) started directing all my personal tax office mail to themselves, scanning it, and sending me E-Mail notifications about them. I then proceeded to ignore those for more than 6 months (did I mention that it was my fault?), assuming that since the company wouldn't have any "real" operations until late in the year that I could deal with the paperwork then.
The result of which was that I got some request for clarification related to the childcare benefit, which by the time I found out about it in mid-December had escalated to a hard deadline the coming January 1st, with re-payments for childcare benefits due in installments the coming February, the first of a few payments being €10k.
I'm pretty sure the last of many reminder letters (although I didn't dig it up) had at that point escalated to include some variant of phrases like "in the name of the monarch!" and "pay, or else!".
Long story short (and looking back at my E-Mail records) the government went from being convinced that I was engaged in defrauding it of something like the median Dutch yearly income in child care benefits, to being very understanding of my situation, and helping me to file all the correct paperwork.
I'm not saying that applied to everyone who got tied up in that bureaucratic meat-grinder, and YMMV, but I for one found the whole process very helpful and understanding.
It probably helped that I wasn't engaged in any "real" fraud. The tax office agreed that if I'd submitted the required documentation on time there wouldn't have been any issue.
Even though the fault of being tardy with that was entirely my own they were very understanding, I didn't have to pay a penny to the tax office, and they went above & beyond in helping me resolve the issue.
I'm sure that if I'd have been engaged in "actual" fraud the consequences would have been very different, but I wanted to chime in with an anecdote-of-one that in my case it was surprisingly easy to go from assumed fraudster to the Dutch tax office bending over backwards to help me, and bend its own rules to help me out.
Another reason is that the majority of Dutch culture/society treats the state-sponsored "independent" media as a church of truth. The majority doesn't question them and all the media did (on purpose?) by repeatedly talking about "the scandal" superficially is just make people tired of hearing about it without knowing what it's actually about.
I think this lays at the core of why they where re-elected (for the 3rd or 4th time).
Of those 1100, a certain percentage would have been placed outside of the custody of their parents with or without this scandal, for just reasons. That's the key here.
There will have been a number of cases where the child's environment would not have changed for the worse, had this affair not happened. Financial troubles tend to exacerbate existing issues. Several of these cases have been reported on in-depth in various (reliable) Dutch media, despite what some people might decry. Those cases are inexcusable, but they do not, according to what is known now at least, form the majority of these 1100 cases, or even a significant part.
Anyone who speaks of '1100 kidnappings' is either naive or pushing some other narrative.
There's zero repercussions or compensation if the state wrongly removes your child.
There's a massive shitshow if they fail to remove your child and "something happens".
So the system tends to trend on the side of taking children into care (traumatic for everyone involved) prematurely.
When you are charged or accused for a crime (say, welfare fraud), social services pretty much always automatically step in if there are children in the household.
Then the "remove child" decision tree really gets going - with social services assuming guilt due to it being their safest option.
Excepting incidents, the current state is that there is no direct link between this affair and an increase in children being placed outside of their family¹. That doesn't mean that this affair didn't contribute to some of those cases, but it certainly didn't mean 1100 children were 'kidnapped'.
1: https://www.cbs.nl/nl-nl/maatwerk/2022/48/actualisatie-uithu...
> There's zero repercussions or compensation if the state wrongly removes your child.
The fraud revolves around small companies ("gastouder kantoren") thinking they found a loop hole in the law. The short version is they would say: "Hey, do your (grand parents) sometimes baby sit your kids? We'll register them as a 'gastouder' company and now the government will pay for the childcare.". The problem was that these "companies" didn't have things like a proper administration which was required to get the government money. The "gastouder kantoreren" would them make up some fictional administration and get caught.
> the reality is much simpler: the Dutch tax office systematically engaged in discrimination, persisted in doing so when this was pointed out to them and made extraordinary efforts to hide their failings rather than to fix it.
You are framing this way too simple. The law which mandated them to come down hard on this fraud was written to combat the "Bulgaren fraude" (i.e. the Bulgarian fraud). In which east Europeans (mainly Bulgarians) would come to the Netherlands, claim benefits they weren't entitled to, get paid and then disappear. There was an public uproar and politicians (like Pieter Omtzigt) demanded stricter laws and hard enforcement of those laws to combat this.
> persisted in doing so when this was pointed out to them and made extraordinary efforts to hide their failings rather than to fix it.
They interpreted the law (which governed their actions) that it required them to come down hard on on fraud. The court initially confirmed (in a ruling) that this was the correct interpenetration of the law. It wasn't until much later that the court reversed on that position.
> If not for a single lawyer (González Pérez) persisting this whole thing would have never been brought to light.
It's worth mentioning her husband ran such a "gastouder kantoor", which was also investigated. You can look up the court documents online. It was quite clear he was running a fraud because for example all the hour sheets for the "gastouders" were the same.
That difference is massive and I don't really understand why if you are aware of it you would even try to make the case that the people committed fraud. They were by and large utterly deceived by parties making money hand over fist and the whole reason that situation existed in the first place was because of a bunch of idiotic legislation that made it illegal for people to arrange for their own child minding services in case both parents worked. To the point that even relatives could no longer legally look after the kids without being at risk of all kinds of nastiness. It's a prime example of the dangers of regulatory capture. That a law written for a special case has side effects orders of magnitude larger than the case that it is written for alone should have been a huge red flag for further execution, but instead of stopping they doubled down on it.
The biggest problem here is that there isn't a single individual that has been held responsible for any of this, all of the parties involved hide behind one another or play the equivalent of 'just following orders'.
Nobody deserves to have their life ruined at the hands of the state, even if there is perceived fraud the remedy should never be worse than that which it is supposed to address.
Throwing shade on Pérez is despicable, she deserves a statue for her persistence especially in light of all of the people that went after her and her family, which you are doing as well here.
As for what constitutes fraud or not: I highly doubt it was possible within the rules as set to not be in violation of one or more of the rules so you could write that up to '100% fraud'. Personally I'm really happy that I never had to make use of any of these services.
Illegal to arrange your own child care? That's simply not true. You can bring your child to your parents, friends or whatever person for day care if you want to. If an accident happens the care takers are most likely not liable. And even if they are liable they are covered by their normal liability insurance (92% of people have a liability insurance in the Netherlands [1]). That is of course unless you start charging money for day care (because then it's a business).
> The biggest problem here is that there isn't a single individual that has been held responsible for any of this, all of the parties involved hide behind one another or play the equivalent of 'just following orders'.
Because everyone is to blame. And they know it and that's why they are not stirring it up. For example hero of the "toeslagenaffaire" Pieter Omtzigt was one of the people who made it possible. He himself was one of the people most vocal after the "Bulgarenfraude" demanding tough laws and crackdowns on such things. And like I said Perez's husband with his "gastouderkantoor" is also to blame. So everyone is happy with this sort of undirected outcry and some reparation payments "resolvement".
> They were by and large utterly deceived by parties making money hand over fist
Sure, they were deceived with a story they could get free money. But people also have a responsibility themselves to know there is no such thing as free money. You can debate on how hard the government should or should not have cracked down on them. But it's also not as black and white as you are framing it.
> Throwing shade on Pérez is despicable, she deserves a statue [...]
You yourself state: "They were by and large utterly deceived by parties making money hand over fist" but stating the fact that Perez's husband ran such a company is "throwing shade on Perez and despicable"? Perhaps instead of saying things are "despicable" and people should get "statues" you should argue why you think it's not a relevant fact.
[1] https://www.consumentenbond.nl/aansprakelijkheidsverzekering...
EDIT: I was wrong about personal liability insurance being required. It is required for your car but not for you as a natural person. Added source stating that 92% of the people in the Netherlands have a personal liability insurance.
Personal liability insurance is not mandatory in the Netherlands [1]. And stating it as such makes me doubt a lot of your other claims as well.
[1] https://www.abnamro.nl/en/personal/insurance/home-insurances... (see FAQ - "Why do I need personal liability insurance" at bottom of page)
[1] https://www.consumentenbond.nl/aansprakelijkheidsverzekering...
This is incorrect. It is only required if you want subsidy.
Not if your parents baby-sit your kids. And that's the scenario where this is all about. People were already bringing their kids to their parents and then someone said: "We found a loophole how you can get free money for your baby-sitting parent who aren't getting paid. We'll just pretend they get paid.".
Your wording makes it quite clear that you have an agenda. I guess your argument would be more believable if you did not use wording like this, just as a hint for future arguments. Besides, not everything possible is also a good idea, that something modern citizens seem to forget.
> Your wording makes it quite clear that you have an agenda.
What agenda would this be exactly, anti-idiocy?
I dont quite see how it would work, but imagine if somehow it could be that I as a parent would simply arrange to pay for childcare services myself, instead of having someone steal my money with the threat of being put in prison, and then petition them to pay a childcare service I have identified.
I cant quite figure out how it could work myself, as this middleman is obviously performing a service that I would never be able to do myself, but maybe smarter heads than mine could prevail, and come up with a societal structure where I can somehow just go pay my nanny, instead of having someone else do it with money they took from me.
I'm an immigrant (UK > NL) and don't see how you can give anyone blame here except for the BD being inept and the gastouderkantoren committing fraud. The systems here are complex and there are a lot of really dumb rules (like a flat on savings based on a 4% interest rate - instead of the earnings being treated as flat income as in other countries).
Taxes are very hard for non-natives, they use a formal style of Dutch which can be hard for people still learning the language.... and learning a language if you have small kids AND WORK(!!) isn't easy (for me it was very easy but I didn't have kids then and I'm in the top 0.1% of intelligence if you believe that such test aren't BS).
Don't get me started on forfaits, they should be illegal. The tax office essentially gets to declare what they believe you could have earned instead of what you actually earned and then they tax you on that.
There's a whole industry of shady companies who help you with tax-savings, for example a very common one in IT is companies who "help" you claim R&D or innovation funding. I've seen that used a lot but never for anything which was actually innovitive.
A recent one is the STAP budget where every citizen can get 1000,- for education and millions went to get rich quick courses run by shady influencers.
Offer free money, people will abuse it, but if you crack down on fraudsters you will probably destroy some innocent people in the process.
Many tax advisors also are dodgy as all fuck.
Especially in the context: to own a house in NL most people need two salaries. Childcare was up until about 2010 easily accessible and affordable to everyone. That was all removed as a cost cutting measure, which is why we had stupid systems like this.
Oh, taxes are very hard for natives as well. The whole tax system doesn't make any sense at all and several elected governments have promised to make it simpler for a long time now.
"People should not even think in these terms." Thanks for telling us all how we are allowed to think.
As a non-Dutchie living in The Netherlands, I don't think the Dutch are inherently are better than Bulgarians.
I'd bet good money that if there was another EU country with a similar social safety net where the median salary was €200k (with correspondingly generous welfare) that a certain segment of the Dutch population would be flocking to it to defraud that government.
I don't understand why it's controversial to point out those basic realities among countries in the EEA. Nobody's claiming that there's a unique genetic link between Bulgarians and fraud, just that the economic incentives involved are all going one way.
1. https://en.wikipedia.org/wiki/List_of_European_countries_by_...
Personally, I blame a longer standing issue of politics designing laws that can not be executed as intended, instead of racism. However, these politics hit minority groups as these are most affected by such legislation. I do believe racism did seriously accelerate the issues here.
In the EU, we have a weberian bureaucracy, this basically compiles to that the institutions such as the belastingdienst will get legislation and execute it as "efficiently" as possible.
However, in recent (±15-20y) politics it is expected that the institutions will actually think / care / morally judge the tasks they are supposed to execute. That just doesn't work, period.
A system such as that is implemented in the US, where a change of political party means the replacement of the bureaucracy / government staffing.
Either we: 1. Start politicizing our institutions and stop asking the bureaucracy to do task X as efficiently as possible.
Or, 2. We make sure that the laws are not written with just the intention in mind, but executionability (and what-if-applied-to-the-extreme) as well.
Unfortunately I don't really see any party talking about institutional change at the moment for the new elections, it's mostly clout chasing in their programmes. That said, the election is for the PS.
Its a problem of incentives and metrics for the social service caseworkers - if they do not act, and "something happens", they get the book thrown at them.
If they do act, and there was no risk, well it was better to be "safe" than "sorry" right?
Basically there's zero negative outcome for the case worker or social service system when they break up a family or take a child into care.
There are negative outcomes if they fail to do so.
There probably are/were cases where the clawing back of previously paid out benefits pushed parents into territory where the circumstances for having their children taken away could develop. However there is little reason to suspect that out of all those who were involved in this and had their children taken away, had those children taken away because of that involvement - i.e. the '1100' number mentioned above.
Also your number 2 is "people should predict the bad outcomes of the laws they write", which just sounds like "get good" but for lawmakers.
would suck for fiji, but be a new utopian era of prosperity here
2) if your supplier has it's admin not in shape (or is comitting fraud), then you are not a fraudster.
3) yes, the law mandated them to track fraud. It also mandates not labelling innocent people as fraudsters.
4) the awesome power of the state comes with great responsibilities. After 3 years of newspaper articles and commissions and debate, it is clear that that bar was not met.
The crux of this issue is the fact that, yes, the state uses heuristics to operate efficiently. And courts have ruled that the state is allowed to do this. The border patrol frisks certain people more often. The tax authorities control self-employed people more often. That's not cool if you are one of those groups, and we would all do well to empathise with them. But as a society we have not found a better solution. The alternative is frisking everyone or having some sort of panopticon, which is not great neither.
The real problem arises when the heuristic becomes a hard rule. And when the heuristic does not lead to an indicator "to be investigated in a neutral way", but to "likely fraudster, interpret anything you find as fraud" or even "known fraudster". Just because there is a statistical correlation doesn't mean that it's always true. This is where the tax authorities went wrong. And this is why they are now compensating people.
In this particular case, the law is written that the liability lies solely with the people applying for the subsidy (which are the parents, they need to register how many hours they wish to apply for, not the daycare providers). Ignorance of the law is no excuse.
"Van Rij admits institutional racism is a problem at Dutch tax office" - https://www.dutchnews.nl/news/2022/05/van-rij-admits-institu...
People might not realize how lucrative this construction was (and still is, AFAIK it still works if you do all your admin correctly). Just like in the rest of the world, it's not uncommon in the Netherlands for grandparents or unemployed relatives to provide daycare for free. But if they register as an official 'gastouder' (guest parent), they can get about €1400/child/month in child care subsidy if the parent(s) are on minimum wage. For context, minimum full time wage is around €1900/month, and 'AOW' (social security for retirees) about €1300/month. Median income in the country is below €3000/worker/month, so grandparents taking care of 2 grandchildren taking the full subsidy would bump them comfortably into the upper half of income earners (all amounts after-tax).
If you have a child in The Netherlands you're entitled to a daycare subsidy. The subsidy is the same whether you send your kid to the daycare down the road, or if you pay a "gastouder kantoren".
It should be obvious how the likely signs of "fraud" in this case are highly correlated with saving the state money overall.
If you're paying one of your parents to take care of your kids that's usually win-win, if they weren't getting that income they might in turn be eligible for other poverty subsidies from the state.
The overhead of paying a "real" daycare is also going to be much higher in administrative overhead etc.
The average daycare worker is also likely to be of an age and inclination (given that they've accepted a job taking care of the kids of strangers) to be otherwise productive elsewhere in the economy, which is less likely to be the case with parents willing to act as a "gastouder kantoren".
So cracking down on this in particular seems to me to be at least partially a case of penny wise pound foolish, which the Dutch are otherwise pretty good about avoiding. This system can absorb a lot of fraud before it would become less beneficial overall than the alternative.
Wtf?
This definitely isn't a universal sentiment.
> Acting on demographics based on statistics is called discrimination and it is a crime
Not if the government does it. E.g. the vast majority of Russians would not be spies, but Ukraine still isn't letting Russians into the country.
“While Dutch criminal gangs have been dubbed the “Mocro [Moroccan] Maffia” and ethnic minorities are overrepresented as crime suspects, Statistics Netherlands researchers say racial background is less important than age, education and socioeconomics.”
Sometimes simple people manage to see the obvious things that the intellectuals are purposefully missing: Europe’s brainless migration strategy accepted anyone with a pulse instead of focusing on skilled workers. These poor and uneducated migrants were isolated in ghettos and turned to crime as an escape.
Now Western Europe’s haunted by “clans” and “troubled youths” while many politicians and the media still haven’t found a way to address the problem without - God forbid - appearing racist.
You call it discrimination and I call it theft by the ruling class against its own people to import foreigners that the ruling class can more easily control in order to arrest control away from the native populations that keep insisting on being in control of their own lives.
It is sad that people don’t seem to want to understand that “immigration” is a weapon of abuse by the ruling class to attack and shatter the “will of the people”.
It has really never dawned on anyone that there is something wrong with the situation when the richest and most powerful people … literal lying, murdering, stealing psychopaths … are all cheering on and support “immigration”, as if anything they support will have any kind of positive impact on regular people?
It’s not very sane to believe such lies by abusive psychopaths in the ruling class. The people who start wars, kill millions, and pillage countries … the very same countries they then cheer on “immigration” from … should be trusted on “immigration” any more than they should be trusted???
It is mind boggling to be to what extent people will go to trust their abusers. The same people who are doing this and who are responsible for things like COVID are also the murdering pillagers that destroyed and murdered Iraqis, among all the other death and destruction they reap. Why are people trusting these psychopaths?
"a foreign (read: Islamic, Polish or Bulgarian) sounding name"
That explains a lot for me, because my last name does sound kind of Polish/Bulgarian and at the same time does not, since my family tree has been of Dutch origin at least from the early 1600s!
I work with our local food bank, and got in a discussion with a family member who was shocked at there being no means testing or proof of need to participate. I'm certain there are people we serve who are over whatever threshold would exist, but the cost of excluding someone who's hungry because they couldn't print a paystub or have access to the right documents is _so_ much higher than any spillage from fraud.
With certain aspects of society, such as with social-saftey-net benefits, when the costs of exclusion impact those teetering on the edge, "non-zero" fraud is absolutely the optimal amount.
1 - https://www.bitsaboutmoney.com/archive/optimal-amount-of-fra...
How did you determine this?
Which doesn't seem sensible.
Generally many "fraud prevention methods" for poor people are more about punishing poor people than actual fraud prevention. Often the cost of enforcement is as high or higher than the cost of the fraud.
Can you describe, step-by-step, how you arrived at the conclusion?
1. There is enough food to go around
2. Determining with 100% accuracy if someone is really entitled to help is unnecessary
The question is assuming the marginal case since it wasn't intended to be trivial. i.e. where the 'amount of food' is limited.
But if there is genuine confusion, since it wasn't spelt out, here's the rephrased question:
"How is this assessment made when food is limited?
This makes sense, thanks for actually writing an easy to follow explanation.
> If fraud rates are small, the cost to everyone is trivial; if fraud rates are zero, the costs get piled onto a few unfortunate souls who couldn't pass the test despite actually needing help.
Can you explain how the second follows the first?
This is the kind of bold claim that needs a citation going to some organization that has studied this claim (maybe the USA socialist party has done one?).
Does the Fraudster Union keep official statistics?
Very poor countries are a very different environment (I lived in one for a couple of decades).
It is quite a lot of work per person, for what 10 or 20€ worth bag? And doesn't even really cover the scenarios which usually result in this.
Some big unexpected cost of money, maybe losing money and the unemployment taking time to process. Failure in that process. So on. Scenarios where technically it might be that person would not even get it yet.
During the debate on the Bulgarian fraud scandal, State Secretary Frans Weekers sustained the Second Chamber’s call for stricter fraud control, but warned that “the good will suffer under the bad”.
[1] https://nl.wikipedia.org/wiki/Toeslagenaffaire#Bulgarenfraud... (Dutch)
People talk about AI enhancing human abilities, and it can obviously do that, but it's really scary how willing people are to trust a system they've been told shouldn't be trusted.
The people they instruct to use it are scared of going against it and being liable or responsible, and don’t have time to verify. Plus “no one gets fired for doing what algorithm says”.
Just imagine a boss is doing that to you, instead of a machine. Giving you wrong facts from time to time on purpose, to see how badly you screw up.
Did you know that they do this with airport scanners?[1] It is very hard to maintain concentration while you scan innocent luggage after innocent luggage. And the chances of encountering a bomb or smuggled weapon is very low, most scanners will work a lifetime without ever seeing one for real. So how can we know that they concentrate instead of just blindly pushing the “ok” button? The system is programmed to preset fake weapons overimposed on the scans a certain percent of time. They are expected to push the “suspicious” button, and then the computer says “you got it, it was me who put the weapon’s image there” and asks the scanning person to check again now with the removed overlay.
It is much easier to filter out one out of ten on average than one out of a million.
How hellish this is depends on the communication and consequences. If your boss does this very rarely, doesn’t tell you about it, and punishes you severly that sound hellish. If they tell you that on average every 5th case you inspect will be innocent, and your task is to filter these out, then it is much harder to get worked up about it.
Not sure also if airport security is a shining example of how much sense these methods make.
The word “algorithm” has become the qualified immunity of non-police world. Seems you can always blame the algorithm and walk away unscathed.
Best we could do was make it configurable on a database so it could be changed and noted that to them "in case".
But at the end of the day, we essentially had a rule that said:
If birth country in (Afghanistan, Serbia, Russia, ...) then add +20 risk points for flagging.
I wonder how powerless software people really are in this context, how much effective oversight others have over us in these matters, where there's sometimes a significant human cost. Just a thought.
It's just being used as a scapegoat to do things they don't want responsibility for.
"Computer says government gets to keep all the money, instructions for an appeal we'll deny right out of the gate are on the back of your printout, have a nice day, NEXT"
>but it's really scary how willing people are to trust a system they've been told shouldn't be trusted.
It's not trust. It's motivated reasoning.
So still following an algorithm, though ran on human brains rather than electronics.
Systems that were meant to detect misuse of the benefits scheme, mistakenly labelled over 20,000 parents as fraudsters.
Amongst the upheaval, little attention was brought to the fact that the tax authority was making use of algorithms to guide its decision-making.
In a report by the Dutch Data Protection Authority, it became clear that a ‘self-learning’ algorithm was used to classify the benefit claims.
When services impact real human beings and false positives have consequences that fall on the least resourced it's important to follow through with human eyeballs and good channels of appeal.
In Australia the parallel mistake was the Robodebt Scheme:
On 29 May 2020, the Morrison Government announced that it would scrap the debt recovery scheme, with 470,000 wrongly-issued debts to be repaid in full.
Initially, the total sum of the repayments was estimated to be A$721 million, however in November 2020 this figure expanded to A$1.2 billion after the Australian government settled a class-action lawsuit before it could go to trial.
In addition to wasting money, it caused all sorts of trauma to already vulnerable people - sadly in some cases leading to suicide (https://en.wikipedia.org/wiki/Robodebt_scheme#Allegations_of...).
My team at GetUp built a tool to automate initiating the process of challenging these requests https://fraudstop.com.au/
I’m aware of GetUp, but this is the first I’ve heard of a product they’ve developed. And that they have a team that’s capable of such developments.
Robodebt was brutal in it's application and general air of intimidation, I wasn't personally affected but I fielded a few calls on behalf of several people I knew who felt crushed by what for them were significant debts (several times any savings they had in situations where they had zero income over enough to barely cover very basic cost of living).
...and holy hell, they charged a 10% recovery fee?
At least in NZ, when our social welfare system recovers an overpayment, there's no interest or fees. The only punitive measures used are criminal prosecution if you overtly committed fraud to a significant extent.
INTERVIEWER: And the cost of it all, Deputy Minister? Seven percent of the gross national produce ...
HELPMANN: I understand this concern on behalf of the tax-payers. People want value for money and a cost-effective service. That is why we always insist on the principle of Information Retrieval Charges. These terrorists are not pulling their weight, and it's absolutely right and fair that those found guilty should pay for their periods of detention and the Information Retrieval Procedures used in their interrogation.
And later on a well-meaning guard advises the protagonist:
Don't fight it, son .... confess
quickly ... If you hold out
too long you could jeopardise your
credit rating.To some extent, a programmer is the natural predator of these individuals, but that's a side matter.
As such, responsibility, the icky bits, those get pushed to The System, wherein accountability dissolves like cotton candy in hot water. The System said these people were frauds, I merely followed my marching orders. The bigger the organization, the more System to serve as a kind of landfill of responsibility, as well as the more inertia preventing real change. Essentially, the avoidance of the labor of thought, consideration, judgment and the like are what set all of this in motion. And people will still follow their GPS off of a cliff because of this.
> More than 700 branch managers were given criminal convictions when faulty accounting software made it look as though money was missing from their sites.
Hard to read without boiling your blood
You can read a report at e.g. https://www.private-eye.co.uk/pictures/special_reports/justi...
As well as Wikipedia of course https://en.wikipedia.org/wiki/British_Post_Office_scandal
https://www.bbc.co.uk/programmes/m000jf7j/episodes/downloads
There is also a book by the same reporter.
I've always been frustrated that you can make millions writing horribly flawed software for government and get away with it. I'm also frustrated that I'm morally not able to pursue that same business model.
Not referring to this incident in particular, but rather, thinking ahead. In a hypothetical scenario, if it turns out that immigrants from a certain country are in fact high-risk for certain kinds of fraud, and the AI automatically picks this up, is a human intervention countering its bias an erroneous one?
It's telling that so much of racism and stereotyping stems from a human desire for convenience and conformity - it is easier to shape your worldview when you've identified enemies and how to classify them. When you add algorithms with the goal of increasing efficiency into the mix, you create a real possibility that it could profile certain populations out of convenience.
Of course, humans do that anyway. And when these humans train AI, they may pass on the bias. But assuming the objective truth does lead to some form of selection discrimination, how does one counter this?
> But assuming the objective truth does lead to some form of selection discrimination, how does one counter this?
The basis of non-discrimination is to judge the individual. Even if 99% of people from some country are known to be fraudsters, you treat an individual from that country the same as an individual from another country. Anything else would be discrimination by definition. If somebody prefers to apply statistical results to the individual, there is no way to do that without discrimination because that is discrimination.
I would argue not just the effectiveness but also the political acceptance of any enforcement and thus the existence of the state.
Benefits as the plaything for politicians are an incredible and seemingly ever growing pile of sometimes contradictory, hindering or impossible to follow rules. Computers as humans will fail to implement them. The best we can do (besides thinking before writing rules) is to surround the rule application with layers of escalation to deal with the fallout and provide feedback paths to fix widespread or egregious failures.
The unwillingness to monitor and fix failures is what made it a scandal. The economies of scale in computer applications made it a big one.
If we're mad that the workers didn't perform a proper investigation or give people due process, then I hardly see why we need to make the predictive model the target of our blame, rather than the human beings with intent and responsibility chose not to treat people with the rights and priveleges that should be afforded to them.
But I can't understand why everyone has decided the "correct" answer to this problem is to ignore reality and then try to delude machines into acting like how we want the world to work is how it actually works.
All while systematically denying the issue, government officials that knowingly lied during public hearings and a complete inability to compensate those that were affected.
Don't drag "wokeness" into this discussion, it has nothing to do with it.
This would be a surprising statement to be true (theres technically evidence for Jesus being the son of god, its just the priors/hypothesis complexity are low/high. And the fact theres more evidence against the hypothesis). If you asked the makers of the law/algorithm what do you think they'd say?
While retail supply of soft drugs is permitted and taxed, production and wholesale supply to those retailers is illegal, which helped drug gangs to become immensely powerful.
This could be resolved by legalising the production and wholesale supply aspects (with great benefits for the taxman), but the political will is lacking. It would also sharply hurt criminals.
The half measures around drug legislation are indicative of how the rest of the country is ran.
Not too long ago, there was also a notorious case of statistical conviction failure [1], where nurse Lucia de Berk was given the life sentence. Initial conviction was based on misinterpreted data which were amplified and extrapolated to a number of suspected murders through bogus statistical analysis.
Just browse a Facebook apartment rental group in Netherlands. Most would say “No internationals!”.
Imagine a society where such blatant discrimination exists, and you will see why such scandals targeting foreigners is no surprise.
Edit: but, in typical Dutch fashion, this essentially innocent statement is worded as rudely as possible.
Human society involves huge numbers, it cant operate efficiently without some sort of automation. Algorithms, (whether simple or complicated versions doesnt matter to first order) have been used since a long time in sectors such as insurance and banking.
When algorithmic decision making is absolutely necessary, regulations, independent control, transparency, accountability etc are various mechanisms to mitigate the risk.
Today there is no excuse to keep dragging our feet about these matters.
Why is that more crucial than another group being disproportionately labeled? Also, why is it so shocking that one group would be disproportionately labeled? Otherwise, randomly audit for fraud and stop pursuing an algorithm.
The focus should be on whether the algorithm is accurate.
Ignoring this horrible Dutch example, I have the feeling that similar policies of trying to "correct" misuse have only managed to leave people in need without their last available support. The UK comes to mind in that respect, lots of similar cases in there (from what I could read).
The question then becomes what is the threshold to open an investigation? An AI represents reality better than a human, but also reminds us that certain classes and races of people are more likely to X Y and Z.
If reality is racist, and there is no such thing as equality or equity, is it ever ok to use an AI for any task at all?
Even while not feeding race data to the AI, it will still be racist because all datapoints should lead to an accurate representation of reality, and there is no such thing as equality or equity. In this example, even without race data, the AI should STILL come to the same conclusions. In fact, I doubt it knew the skin color or ethnicity of the names of the people it was analyzing.
Where the Dutch fucked up was jumping to conclusions and labelling without cause or investigation 20,000 people as fraudsters.
BUT, EVEN IF THEY CONDUCTED INVESTIGATIONS ON ALL 20,000, and they were guilty, does that make it any more right or less wrong?
I posit no. It's still illegal and wrong.
AI can only represent reality, reality reflects uncomfortable truths. AI has to be racist, or it isn't doing its job. The world is not equitable, all races have difference cultures and values and work ethics.
Which I posit means that even if the AI is correct, and even if those 20,000 committed fraud, its still an illegal prosecution as there was never a legitimate predicate to investigate in the first place.
AI is racist, even if it represents reality, and therefore cannot be used to initiate an investigation.
To me, a more interesting question is whether AI should be allowed to assist police in investigations, as even with a predicate for opening an investigation, the AI is still racist/sexist/classiest etc, and is going to help put people in jail based on racist classest sexist biases.
An example is a middle class white man that shares few traits with the average criminal, but is nonetheless a criminal. VS a black low income man that shares many traits with other criminals.
In this example, if an investigative predicate was found for both, and an AI was used as an assistant to the police, the police would be looking at the low income guy with more scrutiny as they would believe that a more likely case to win. Theoretically. Even though we know them to be guilty, the police don't know that, and the AI don't know that, but the AI is allowing for more scrutiny and conviction against one of the two guilty parties.
TLDR; AI is fucked. Shut it all down.
[citations so very needed]
This is in the context of babies. Babies are expensive. Presumably, immigrants NEED more help than the general population with friends and family. Presumably they HAVE more babies.
Put the two together. Need more assistance, have more babies = more fraud.
It's why I also said it has nothing to do with race. Class and income would lead AI to the same conclusions.
This is precisely the kind of broken reasoning that allowed this system to be created in the first place.
In reality immigrants do not commit more fraud. Besides that many of those that ended up in the dragnet were second or even third generation and as Dutch as I am other than that maybe their skin has a slightly different color or that their last name or first name shows them to have a reasonably recent immigrant in their family tree.
Less fraud then? The exact same amount? That would be surprising, given wildly different circumstances.
This scandal seems to be the direct result of overcompensation for the lax enforcement of widespread benefits fraud by Polish (not-actually-)immigrants. In general, doesn't ik make sense to scrutinize recent immigrants when handing out free money? That has little to do with skin color.
Hey congrats, you're not racist, you're xenophobic!
A good system scrutinises everyone equally when handing out free money. A system that scrutinises immigrants more thoroughly is a bad system that will fail to catch fraud when it's committed by people born there, just because investigative resources are finite, so a focus in one area necessarily reduces investigations elsewhere.
Also, let me tell you an anecdote to illustrate why it's the people born there you really want to keep an eye on.
In the early 2000s, I worked in my country's social welfare department, interacting with customers and ensuring they had full and correct entitlement.
On my caseload were a notorious couple, who knew more about social security law than we did.
Our social security law pays a slightly lower rate of benefit per person when you're in a "relationship in the nature of marriage" compared to the rate you'd receive single.
And this couple had spotted a potential loophole through close reading of the Social Security Act 1964.
Sure, they had kids together, and lived together, and had been married for years. But, what if, what if you claimed that the wife had just realised she was a lesbian, was now in a committed relationship with another woman, and only living in the house with her husband to look after the kids?
That'd be two single benefits right there right?
(The answer was no, because the definition of "relationship in the nature of marriage" didn't require you to be having sex, because, plenty of married people don't)
These people, they're born and bred, and they'd grown up in intergenerational poverty, so they knew the laws very well.
A migrant would not be able to compete with them at all in a benefit fraud contest, they lack the institutional knowledge.
TL;DR - everyone who is poor and wants to be less grindingly poor has motivation to commit benefit fraud.
So what makes you think migrants are more likely to act on that motivation than locals?
Do you genuinely believe desperately poor Dutch people would abstain out of honour or national pride or, I dunno, something else equally implausible?
Just my anecdote, of course.
Good in what way? Cost-effective or fair? I grant you the second point.
> everyone who is poor and wants to be less grindingly poor has motivation to commit benefit fraud.
Sure. But being a small country with (what uses to be) a generous welfare system obvious leads to relatively many (fake) immigrants with fraudulent intentions. Just a numbers game. But I guess you could call it xenophobic to take that into consideration when semi-randomly selecting people for extra checking.
But then it really depends on the nature of the fraud you're mentioning.
Is it people lying about their qualifications for entitlement?
Not declaring income?
Discrimination is discrimination. Its simply not allowed.
Yet if all roads lead to an accurate picture of reality, and reality is not equitable,
How is any investigation ever to be opened?
A neighbor is suspicious and watches an immigrant family like a hawk and reports things to bylaw and police? It's discriminatory in nature, yet legal because a human observed a predicate.
I think that is what we have to keep in the future. A human must witness the predicate for an investigation. Even that is flawed. But AI is on a whole other level.