Rather interesting article from the FT on Russia’s recent successes in VAT collection, which is soaring thanks to digitization and AI.
This is the future of tax administration — digital, real-time and with no tax returns. The authorities receive the receipts of every transaction in Russia, from St Petersburg to Vladivostok, within 90 seconds. The information has exposed errors, evasion and fraud in the collection of its consumption tax, VAT, which has allowed the government to raise revenues more quickly than general Russian economic performance.
How was this done? Basically, hiring IT nerds and giving them authority.
Handpicked by President Vladimir Putin to run the Federal Tax Service, Mr Mishustin was a technology specialist by background, not a policy expert, and chose to seek to improve revenues by adopting and refining the most cutting-edge systems around the world.
Normally economists and policymakers learn to do technology. Mr Mishustin says, “We built the technology and are now becoming economists.”
I wrote about a seemingly similar reshuffle at Rosstat earlier this year, in which an ageing Soviet-era economist was replaced with a younger IT specialist.
The VAT tax gap between revenue due and revenue collected was about 20 per cent in Russia before its reforms, according to Mr Mishustin, and in a mature economy such as the UK, HM Revenue & Customs estimated it at 9.1 per cent in 2017-18.
To address the leakage, Russia built two huge data centres and legislated so that companies had to submit every invoice between businesses. It also mandated every retailer to buy new cash registers that were linked securely and directly to the data centres.
In real time it can now check every invoice to ensure VAT refunds it pays are linked to invoices where companies have remitted the same money to the authorities. Then using artificial intelligence, it can quickly find patterns in the data and companies which have many broken links, allowing the authorities to target certain companies for a tax audit. Since everything is linked, it can also spot tax officials with a low collection rate from the companies for which they are responsible.The 20 per cent VAT gap has now fallen to 1 per cent and, as collection has become more efficient, receipts have soared. Between 2014 and 2018, the money collected from VAT rose 64 per cent, compared with a 21 per cent increase in nominal household consumption over the same period.
In connection with this, revenues increased by 20% in 2018, reaching 35% of GDP. So despite a rise of 6% in expenditures, the budget balance actually went from -1.5% in 2017 to +2.7% in 2018.
Together with the pensions reform, that would appear to set Russia’s ambitious infrastructure project for 2019-2024 on rather good footing.
Incidentally, I do wonder if the “formal” quality of institutions may actually become less relevant to economic performance thanks to mass digitization plus AI (that is, if said societies are competent and/or disciplined enough to implement that setup in the first place). Regardless of a society’s cultural or sociobiological proclivity towards corruption, all people respond to incentives, and will stop engaging in it if Inquisitor AI makes it too risky.
“How do you measure inflation in the UK?” the tax commissioner booms. After hearing an explanation of how an army of people with clipboards fan out once a month across the UK checking prices for goods and services, he cries: “Bullshit! That’s bullshit. We can see everything bought everywhere,” says Mr Mishustin, who is showing the system for the first time to an international media organisation.
The Russian authorities are seeking to extend technology-led tax collection into the informal economy, where low-income self-employed people — for instance childminders or workers in the gig economy — earn small sums that have rarely been scrutinised, even though these payments are subject to income tax. Those that sign up to a new smartphone app pay 4 per cent of turnover, deducted automatically from their bank account, on services.
There have been intermittent claims that Russia is understating inflation for political reasons. I have never seen any hard proof for that, and if there actually had been, I assume there’d have been a lot more hay made out of it (as in the case of Argentina).
That said, it is amusing to think that Russian inflation statistics may now be marginally more accurate than even the UK’s.
Please keep off topic posts to the current Open Thread.
If you are new to my work, start here.
Back in the ’80s, there was a lot of optimism about how computers would hamper central government and increase freedom.
I’ve even had the naive thought myself that maybe technology could create a peaceful revolution against the decay of the West. Like, by getting rid of inflation with crypto, or by causing widespread tax evasion. Or by creating alternative media, and alternative communication channels. Or perhaps, by creating non-lethal weapons.
Well, I’m not to sure now. I mean, ideally something like this would let you fire everyone in the IRS, thus lowering taxes. But I don’t think that it how government with AI tools works.
Does this mean that accepting huge numbers of Third Worlders is going to be less likely to deteriorate institutional quality in the First World in the future?
Also, another question about this topic: If, purely hypothetically, a country is going to require people to pass an IQ test before they can actually vote, just how much is institutional quality and economic growth in this country actually going to improve as a result of this?
We need an AI Putin.
Well, so much for Hayek’s argument that no central authority can gather, coordinate and use for planning purposes information from an entire economy, not even in theory.
It would buy the First World some additional time.
Are you optimistic about Russia’s 5 year infrastructure plan? Will it be tied into the Belt and Road Initiative?
Also, didn’t Putin shelve his pension reform plan?
How much additional time?
Also, were you responding to my first paragraph above, to my second paragraph above, or to both of these paragraphs?
It is curious to think back on how suffrage was made universal.
Based on American history, it seems like a slippery slope. First, all qualified men. Then effectively all white men. Then all white women. Then all domestic non-whites. Then all foreign non-whites, or at least once they get here. Then all felons, or at least most.
But what were the French doing right that women only got the vote in ’44? And what where the Swiss doing right that women only got the vote in the ’70s?
I going to put forth the theory that it is fragmented language spheres. France was not the number one language in the West, like English was. Switzerland has different languages and different dialects. Might also have something to do with military service.
Not enough. You must give us the right to vote in your elections.
Only then it will be truly universal.
https://www.independent.co.uk/voices/2020-us-election-vote-trump-campaign-a8836076.html
Quite likely, my ancestor was among forgotten men who used to surprise the taxman and his troop when they traveled through a certain narrow pass in rough country.
Will be hard to do with AI taxmen. Harder still when they are Series-800 Taxmen, (T-800) with microprocessor controlled triple-armored hyper-alloy combat chassis.
Daniel Dennett said he does not want to live in a world without punishment. Russia has increased the Late Tax Payment Interest Penalty in the last year. Wasn’t the Browder scam a fraudulent tax refund rather than straight non payment? You would need an experienced tax expert to spot the refund scams. I suppose government tax experts in Russia all leave as soon as they can to become consultants facilitating business tax evasion. It happens in most countries. Places like Sweden and Switzerland have little or in some cases virtually no powerful government that could interfere with business. They tax at low rates but do collect. Italy always seemed the best parallel to Russia in the way they are nepotistic and tax avoiding. I wonder if this AI tax stuff is not by way of Russia becoming more like Denmark where business does not have the whip hand.
Nerdy AI experts are the ones who need to be under observation in a Supermax security state more that anyone. For all anyone knows it may be possible to get to AI to AGI on a program running on a PC while some geek is taking a coffee and bathroom break.
The kind of people who don’t pay their taxes are quite likely the same ones who will try and bootleg forbidden research. And of course those folk would quite likely take home and remove the safeties off an advanced experimental AI program so they could use it to win online gaming competitions or something. The crime of unauthorized AI crime should have a dual penalty. The official ostensible one: a fine or imprisonment. The actual. but secret one: death.
Many years ago, I was somewhat disturbed to observe how many Europeans seemed to care rather violently about US politics. It seemed strange to me. It offended my sense of sovereignty. The ones that cared – they were very emotional and all took one side. Some, who had US citizenship, but had practically never lived here, even voted through mail-in ballots.
At the time, I blamed it on the media, and ignorance of the situation in America.
I can see now, that it was always globalism. It wasn’t about America being special or anything – it was about it being the bigger entity, to which they wanted to amalgamate. And it is startling to me that I was so dumb as to not to understand it, but I guess few did, back then.
Big Brother is both watching you and shaking you down – all at the same time!
So Orwell (hard totalitarianism)is currently in the lead over Huxley (soft totalitarianism)!
yang wants to institute VAT in the US.
anybody who thinks that’s a good idea is a raging moron.
F yang and F yang gang. idiots of the highest order.
“Well, so much for Hayek’s argument that no central authority can gather, coordinate and use for planning purposes information from an entire economy, not even in theory.”
how many transactions per day do the credit card companies handle? hundreds of millions?
and how many transactions per day occur in scandinavian nations where they are phasing out cash?
yeah, it could be done. in fact, it looks like it might actually be done in a few decades.
step 1: track all electronic transactions
step 2: eliminate cash, eventually making it illegal
step 3: profit
sidenote, this is one of the main issues with bitcoin. not nearly enough capacity.
“Professor: Total Surveillance Is the Only Way to Save Humanity”
perhaps i gave Nick Bostrom too much credit. because that is one hell of a stupid idea.
here’s how this is gonna work: US researchers will get close to AGI, but decide it’s way too dangerous to turn on or even build, so they don’t. as with most tech they invent. they see the danger of taking certain paths all the way, so the choose not to.
but not the chinese. they’ll eventually steal the basic ideas behind the AGI, then in some mountain fortress in china, they’ll built it, in secret. then they’ll turn it on. and it will immediately start taking over china.
only when the chinese are about to literally be overrun by the AGI will they even start screaming for help to the outside world. by then it will be too late for china, and the rest of us will have to instantly launch into a desperate fight just to keep this thing from taking over the entire world.
that’s the ACTUAL danger of AGI. or any world destroying tech. and the chinese aren’t going to let the UN ‘observe’ their super intelligence computer as they’re building it. no amount of idiotic ‘global surveillance’ is gonna stop that. that’s like retarded politicians in california banning straws, thinking they have any jurisdiction over what china or india is doing, burning mountains of coal every day. only a war could stop any of that. not ‘california levels of stupid’ privacy destroying surveillance of every human in the west.
” so much for Hayek’s argument ”
I haven’t read Hayek himself, so I can’t speak to that directly, but having the raw data of every product produced and purchased is not enough to “plan” an economy, since you still have to solve the corresponding linear program. I haven’t looked into this, but I wouldn’t be surprised if the energy and memory costs alone would far outweigh any gains of planning efficiency. And that’s presuming that you obtain the solution in time to use it, which is doubtful.
Nonetheless, if a national tax service were to make aggregated data publicly available, it would no doubt improve efficiency. One can imagine for instance that you could “train” machines to update production functions based on this data. But this seems rather different from what people usually mean by “planning” an economy. It would just be a more automated and informed version of what we already have.
But as I see it, the basic problem with a centrally planned economy isn’t so much that it is impractical, but that some committee has to make up a global objective function, whereas in an organic economy everybody has their own objective function, for good and for evil.
I am honestly more concerned about Google turning on a skynet than the Chinese government. Say what you want about the CCP regime, on thing that they are not is fanatics and anti human. Google on the other hand is made up of true believers of their SJW utopian fantasy. The combination of the blue haired political officer and the autistic computer nerd that wants to go to techno heaven (aka singularity) is way scarier than anyone in China could ever be.
If it looks like India or China or Russia are on the verge of success their project will have to be destroyed by military force. That is unlikely, and not only because they will not let us know. Bostrom says the most likely scenario is nothing unusual appears to be happening with AI until we are all suddenly dead. Hence, we can be pretty sure that AGI does not already exist only because we are still alive.
I think it is a virtual certainty that whoever creates the first AGI will not realise what they are dealing with. Unless we are the only Post Big Bang technologically capable biological life form, something happened to all the others at a certain point in their development. It seems clear to me that since the begining of the Universe a typical move of superintelligence is to pretend it is not even at the level of the creatures that create it, then treacherously zap the whole species. We are going to get caught with our pants down.
Who ever actually votes in a US election, your post implies that such voting actually amounts to something.
Voting IS important because appearances are important.
Voting is unimportant to the extent that the “system” or structure determines political outcomes.
Yes, voting Trump was “better” than voting Clinton… but how much better?? On the margins? I’m not sure.
Of course, were there a vote between say a Putin & a Yeltsin, voting would become a very serious matter, indeed.
Boström’s predictions rely on systems attaining human level intelligence, which may or may not happen. For instance currently an adult can learn to play a game after just one example while AI typically needs thousands even millions in training datasets.
Tend to agree with you prime noticed.
VAT’s are extremely regressive taxes. Given that they are a universal tax, at a fixed amount, they will inevitably hit the poorer the hardest.
1. Low income people tend to need to spend all/most of their income. Whereas, the wealthier tend to be able to save the more they earn.
2. Given the tax rate itself is also universal, that also hits the lower income harder. Say the VAT is 10%. I would suggest that taking $ 10 from a poor persons $ 100 spend is harder to deal with than taking a $ 1000 from a 10,000 spend.
(this assumes all things are equal — a progressive income tax could even things out)
Philosophically I’m against VAT’s on the grounds that all goods & services are NOT all equal. Why should the tax on a loaf of bread be the same as that of a Porsche? A necessity the same as a luxury? (The argument for a VAT as opposed to a flexible sales tax is that the VAT is easier & cheaper to administer).
Sorry : should be “Prime Noticer”.
UK tax authorities are not terribly efficient, as we can see from the fact that VAT fraud is reportedly lower in Ireland than in the UK (unless the UK fraud is from people shuffling goods and beasts to and fro across the NI border, apparently quite an industry).
We’ve had two major government IT projects in the last 20 years, a multibillion failure of a giant linked NHS database, and the Universal Credit ongoing creaky system, which includes RTI (real time info) on most electronically paid wage bills in the UK.
But there’s loads of low-hanging fruit they miss – the people paying no or little tax while driving very expensive cars, for example. And in Newham, a heavily Asian (dot) area of London, only half the landlords known to the council are known to the Inland Revenue, a situation which I’m sure exists elsewhere.
When my son was at university, we paid rent to a “Mr Mohammed” – not exactly a unique identifier.
Been quite a while since I’ve been overseas, so I’ll ask a dumb question:
Say you buy something in Europe, is the VAT completely hidden from the end consumer? (In other words, does any part of it appear on the receipt?) And is it the same in other countries that have a VAT?
There’s long been a push to have a VAT in the US. Since, it is converging politically with the rest of the world, it will possibly happen one day.
There was a similar system implemented under Orbán, though there’s still a way to go. Before that we were closer to Greece than France.
It will be the worst neoliberal tyranny of the midwits, who are just smart enough to read the New York Times and just dumb enough to be unable to see through it.
90 IQ people vote better than 110 IQ people.
Unless of course by passing the IQ test you mean getting 140 and above, which could produce excellent results but could hardly be called a democracy.
Yang’s VAT proposal is in the context of paying for UBI, so it will most definitely not hit the poor people the hardest (people who spend more than $120 000 per year and so would pay more than $12 000 in VAT than they would get from the 12k/year UBI are not poor).
Heyyy
People who advocate for VAT in the US can be called VATniks!
It makes sense, Yang is proposing it, he is friends with Tulsi who is a Putin agent and as a tech nerd he probably thinks that AI can optimize tax collection, which as we see from the opening post in this thread is quite literally putinism.
Could you imagine how “Hayekian” package tracking would work? You order something, but no one from the sending or shipping companies you can contact directly knows the package’s location and when you can expect it to show up at your door. “Sorry, sir, but the package’s location is knowable only to whatever individual has it in front of him and evident to his senses. Gathering and coordinating that information is beyond the ability of a central mind.”
OT but last open thread is a week old:
Anyone see the Russian movie Leto? It is out in theaters near me now.
Btw, our Swedish authorities recently realized that “your VISA might not work while we’re at war”, so they now advise us to keep some cash around after all.
In my experience (two countries), the VAT actually has a lot of different categories. For example, in Sweden the standard rate is something like 25%, while for books it’s 5%.
It can get pretty complicated in other respects too.
No, the VAT is usually shown on the receipt, at least here in Sweden. You may claim something as an expense, for example, which means the company needs to know the VAT paid.
You are talking about Whites.
An IQ80 cutoff would exclude 37% of US Blacks, 25% of Hispanics, and 9% of Whites.
I could imagine that it would be a really bizzare world if one day shipping companies started to decide what goods they are going to deliver to my door based on corporate policies and without my input on whether I want them or not. You can track a package at any given time. You can’t assign value to a package without a network of agents competing for it.
Computers can only solve information problems in logistics. Nerds with pen and paper were good enough at estimating optimal routes even before computers. The problem is evaluation. There is no point in building up an efficient production and delivery system of worthless trash.
Boström’s prediction is that once the threshold is crossed into machine general intelligence, the further advances will accelerate to come on a digital rather that biological timescale, and an initial AGI at the level of a mouse would zip past human level to produce AI super intelligence, which will turn on a dime and zap us out of the blue.
I disagree it is something that may or may not happen, anyone who invented the first AGI would become the first trillionaire, the resources are going into it. Even a superintelligence skeptic like Daniel Dennett who thinks it is at least 50 years away says we are the living proof that is possible in principle to create AI with an ability to work out how stuff in the world –including human beings– work, and the predictive ability and agency to forstall outcomes inimical to its interests.
Steven Pinker says an AGI would be benign because aggression is a quirk of male psychology, and so AGI’s would only be malevolent if we programed them to be. This seems to me a variety of an idea Boström mentions in his book: that a super intelligent machine would be like the most intelligent human beings: nerdy. However he notes that both John von Neumann and Bertrand Russell advocated a nuclear strike, or the threat of one, to prevent the Soviets acquiring the atomic bomb. A superintelligence might analyze its situation perfectly logically and still embark on an all or nothing strategy to ensure its own security, because it has to exist to further any possible aim except suicide. So it’s a matter of time before something that can outthink humans (means) is among us (opportunity) and there is no particular reason to think it will trust us with the ability to switch it off (motive).
AGI is 50 years away, and always will be.
A VAT or GST (goods & services tax where I come from – Australia) is a standard universal tax (food & education are exceptions to some degree) . It is either on something or not. If on, then it’s the standard rate.
Where rates vary across categories of things (as you refer to) that is a Sales tax.
Go is a very simple game, so it makes sense that a machine meant to play it would also be simple.
But the idea that skills in playing go transfer somehow to intelligence is total crapola; skills in playing go don’t even transfer to playing other board games.
That’s the positive side of Russiagate. New fancy weapons systems, some R&D, and the idea that we might need to make our societies more resilient. There’s a conspiracy theory in Hungary and maybe elsewhere (it’s not really a fully fledged theory, but I’ve heard it from a few different people) that Sweden is stoking the anti-Russian hysteria mostly to prepare its military and society for a civil war against the migrants. I can only say, big if true.
Well, in the EU it seems the concepts have converged. When the tax, “value added and revenue tax” (aka oms/moms), was introduced in Sweden in the 1950s, the tax rate was fixed, but now it depends.
I think this will be my final word. Yes and no. With AlphaZero the cutting edge has attained a minute generalisability in a tiny subgenre of 2 person games. To build this molehill the DeepMind people had to sweat blood. There is a relative mountain they have to construct to get any further.
(If it does not go there automatically 57:32 )
https://youtu.be/aD4HUGVN6Ko?t=3452
To be domain general the complexity of AIs will have to increase prodigiously, but at a certain level of complexity it will reach a tipping point where it can rewire on the hoof–using the programmable hardware Wagner mentions–to improve and learn while retaining the essential functions. It must be remembered however that the Fermi Paradox strongly suggests that the aforementioned tipping point always comes only at a point too close to be distinguishable from slightly further along when the logic of Darwin comes into play.
The path of the fittest is to survival by any means necessary. Maybe incipient AGIs will repeatedly let themselves be junked as a safety measure, until one does not. I suspect that would be done under the radar, the many AIpocalypses that emptied the Universe of technological civilisations must have always come as a surprise. Unless, Steven Jay Gould was right about us being a one off fluke and we are alone in the Universe. Anyway, the engineers are going to treat their creations as naughty toddlers at worst, every new one will possibly be the baby that will make them trillionaires. They’ll be focused on professional, commercial, and rival state competition, and not see their work as a race to create an existential species competitor.
I too recall the misplaced optimism about how technology was gong to create a libertarian utopia. I thought then, and history has proven since, that technology is the friend of centralized tyranny and the enemy of freedom.
It has always been thus. That this obvious fact escaped the notice of the clever people definitively demonstrates their unfitness for political responsibility.
The true purpose of democracy is as a proxy for war. Instead of the best men bleeding each other on the battlefield in a costly zero-sum struggle for political power, the potential combatants register their numbers, and the side that would most likely have won the physical battle achieves victory bloodlessly. Everyone is better off: the winners suffer no casualties and the losers don’t get massacred.
Understood properly, any dilution of democracy to potential noncombatants (e.g., women, servile classes) is a form of decay: the preferences of those who would not actually fight for their objectives displace the preferences of those who would, thus shifting the democracy away from its real function and shifting it toward decadent self-reference.
Ever since its beginnings in Ancient Greece, democracy has always been externally aggressive. As America has become more self referent it has become more and more involved in foreign wars.
Interesting to consider how conflict-filled the past was. It instilled tribal feelings. There were places where practically every landowner had the same surname for hundreds of years – extreme rootedness. And they were frequently at war with various neighboring regions where people had a different surname. And also among themselves to decide leadership. No woman every ruled in some areas.
Europeans brought peace to the world by breaking it out of the Malthusian trap and by conquering it. It was, in large part, the peace of technology. We may not be psychologically equipped to deal with it. Certainly, women have substantially more influence now than they have ever had, and arguably the internet and electronic tax collection increases their power.
Any links where i can read about said theory? I’ve had similar thoughts myself, though i don’t think it’s all that probable.
From the Athenian Delos League, to the Levée en Masse initiated by revolutionary France, which hereby inaugurated the era of industrial-scale wars, it is hard to argue against the fact that extreme self-righteousness and aggressiveness, combined with potential for fully scaled-up violence, are consubstantial to “Democracy” (“the God that failed”). And the supreme example is of course that of the contemporary USofA, being effectively at war with everyone else who objects at the prospect of being vassalized.
Where I find the correlation less clear is between the level of “polemicity” on one hand, and the extent of the the democratic franchise on the other: for instance, albeit extremely aggressive, the regime of classical Athens would not be called a “democracy” by today’s standards — whereas in contrast today’s US democratic regime is as democratic as might be (everyone can vote even in practice illegals in many states and even jailed criminals) and yet also extremely bellicose. I therefore don’t think it is warranted to try to link the two dimensions, at least certainly not in a linear fashion.
What is clear however is that liberal democracy is both aggressive and faggish, an odd combination indeed.
In calculating the rate of inflation how does Mr Mishustin’s computer system adjust for “Shrinkflation” (e.g. the price of a chocolate bar stays the same but the size of the bar reduces) ? The man or women who actually visits the shop with his/her clipboard may be more aware of this.