Detecting Money Laundering | Hacker News Reader

Detecting Money Laundering

conf.startup.ml

89 pointsarshakn89 comments

The direct and indirect cost money laundering regulation has reached a point where it far outweighs the benefits. Thousands of clients have to jump through cumbersome hoops just to increase the chances of catching one single money launderer by 1% or so. Even the IRS estimates the implementation costs of Fatca at 100 billion worldwide. In reality, it is probably much higher and far exceeds the amount of additional tax 'income' it generates. We really need to start to discuss the negative externalities of overzealous regulations.

Citations needed.

It's not like money laundering is a victimless crime.

Quite contrary, only money is laundered if it comes from illegal practices, which tend to have victims.

Also, what do you think happens if you relax money laundering regulation?

The problem is that the set of things that are illegal in some country at this particular time, and the set of things that are "rationally definitely wrong everywhere, for all time" are never quite overlapping perfectly.

Sex slavery? Arguably rationally wrong for all time and space.

Drugs and prostitution? Debatable.

For example, regulating drugs and treating addiction as a medical problem is probably much cheaper than waging an unwinnable drug war PLUS trying to find all the laundered money coming from it PLUS forcing all the "good actors" (non drug industry people) to comply with the regulations at no trivial expense.

This is true for every economy that refuses to go away despite being wholly illegal. In those cases it is (IMHO) best to just regulate it... Same as what happened with alcohol after Prohibition. Nobody's making moonshine and killing bootlegger opponents anymore, and alcoholism is a manageable medical problem.

So your argument against tracking and enforcing money laundering laws is that laws in general are silly because not everyone agrees with them?

No, my argument is that they are a waste (literally, they cost more to implement than the money that is recoup'd from "illegal" activities).

Certain laws are truly silly, however. I am allowed to maintain that opinion. I may also be anti-authoritarian in general, and do not believe in the vast majority of consensual crimes.

> literally, they cost more to implement than the money that is recoup'd from "illegal" activities

I'd presume there is a deterrent as well as a direct effect on reducing money laundering; I'd expect to see more money laundering if there were no checks.

Quite how it would be measured is another question.

There are a few basic problems with money laundering regulation:

1. The state wants it done because the intelligence is immensely valuable, not just from a law enforcement perspective but also from an international realpolitik / Great Game perspective (e.g. think tracking Putin's billions for the US, understanding flow of US Treasuries). This means the state tends to punish missing even a single money laundering lapse severely (which has the side benefit of lining their coffers).

2. Banks and other businesses don't want to do it because it's entirely a cost center to them. This incentivizes them to do the bare minimum to give the appearance of compliance to avoid prosecution while rarely doing the work effectively in practice. Business AML departments also have the problem of often working at cross purpose with the firm's other, frequently more prestigious and connected coworkers (think Goldman's main guy in Asia pushing through the 1MDB transactions).

3. The law pertaining to cooperation across institutions, section 314(b) of the USA PATRIOT Act, is voluntary. Money laundering almost always occurs across multiple institutions and frequently across multiple jurisdictions, many of which are stingy with intelligence or even purposeully opaque (Panama, Singapore, the Caymans, etc). This results in fragmentary and incomplete investigations and intelligence gathering, leading to an enormous amount of wasted and duplicated effort.

If the countries of the world were serious about money laundering, they would focus on real cooperation and beneficial ownership requirements. Unfortunately the world's intelligence services all work against each other at one level or another, and they benefit from financial nontransparency (funding operations, obscuring flow of funds e.g. the US Treasury refusing until recently to publish info on Saudi ownership of Treasury bonds). Businesses are swept up in the game and even the few that genuinely seem to want to do the right thing find it impossible to overcome the hypocrisy of the game. As many commenters have noted below, money laundering is not a victimless crime. It aids human traffickers, arms dealers, and drug lords. Unfortunately it also benefits states, and businesses get caught in the crossfire.

With the possible rise of cryptocurrencies its only going to get worse. I believe governments should back off from income and profit taxing, its going to become more futile then the war on drugs.

Governments should be focusing their taxes and revenue on tangible things that cannot be hidden or evaded, like property. This would solve so many problems in places like London or Vancouver where foreign money floods the market. Why waste time tracking endless trails of shell companies when you can just tax the wealth when it appears in public?

Plus I see income and revenue as non-zero sum, one man or company making more money doesn't preclude someone else from creating wealth. Taxes on this are just a friction on the economy. But property is zero sum, every piece of coastline bought up by rich people as 4th or 5th homes means entire neighborhoods of normal families can only see water when they are on their 2 weeks of vacation.

>Why waste time tracking endless trails of shell companies when you can just tax the wealth when it appears in public?

Because the wealth should be taxed in the place (country state, region etc.) where it was made (and whose resources were used in the process) and not where it appears in public.

Tell that to the IRS, which insists that even if you're working in a foreign country, for a foreign company, on foreign projects that they still deserve their cut.

The US is one of the only countries on earth that follow this practice.

So does Canada.

Taxing a non-resident citizen's foreign income? No it doesn't. Once you declare yourself a non-resident, you don't even have to file a return anymore.

If the US doesn't acknowledge non-resident status, then you're right, Canada isn't like the US. However, as a Canadian who was "working in a foreign country, for a foreign company, on foreign projects" (in the US) for half a year, I can guarantee you that I had to pay both US and Canadian taxes on that income, and I very thoroughly investigated the case because in spite of all the treaties, I ended up double-taxed for a few thousand dollars' worth.

Yes, the special case in the US tax code is that non-residency does not excuse you from paying taxes to the IRS (although I believe there is a non-residency tax deduction on the first $80k or so).

for half a year

That's the key part. I don't know the exact legalese, but you probably still qualified as a Canadian resident.

Once you're no longer regarded as a Canadian resident, you no longer owe Revenue Canada anything. It's a clean break.

In the US, no matter how long you live outside the country, as a citizen you still need to file a return and pay taxes on foreign income.

Yep. But it's not just time, to be a "non-resident for tax purposes", you have to close all of your Canadian bank accounts, renounce your government-provided health insurance, etc. I suppose it's not an unreasonable system.

This might sound like a pedantic correction, but IRS people I've spoken to are quite adamant about this: the people of the United States have, through their Congressional representatives, decided that they deserve a share of fellow citizens' foreign-sourced income.

The IRS has no discretion in this matter; they just deliver the bill.

Ease of collection is just one of the goals of a tax system. Others include fairness, and not distorting the underlying economy more than necessary. Taxing only tangible things might make collection easy, but may not be fair (it'll generally tax higher income people at a lower rate than lower-income people), and may also create artificial incentives to engage in activities that result in non-taxable wealth rather than taxable wealth.

It will only tax higher income people at a lower rate if those higher income people don't receive any tangible benefits from high income. That's a good thing - the person who receives a lot from society will pay high taxes, while the person who receives a little will pay low taxes.

In contrast, with income taxes, the person who produces a lot for society will pay high taxes.

It's well accepted by economists that a consumption tax is the least distortionary tax.

Your argument depends on a stipulation of the definition of "fairness" where the persons most easily associated with production are entitled to the most favorable tax treatment. Without that assumption, all you're saying is that just as there is a suboptimal tax system that is least expensive to collect, so too there's a suboptimal tax system that is least distortionary to the economy.

There are notions of fairness that aren't premised on the idea that tax rates aren't a point score system for rating people, and it can be rational to accept some degree of distortion so long as it's not so powerful that it prevents the economy from serving its purpose (and so long as that distortion is buying us something).

I'm making two arguments. One is based on the economic concept of distortion - people making choices they wouldn't make absent the taxes. That's a strictly positive claim and it's not based on cost to collect.

The other is the normative one. My premise is that people should pay in proportion to the benefit they receive from society. Do you disagree with this claim?

I agree with your positive claim. Some tax systems are more distortionary than others. I think it's likely that consumption taxes are less distortionary than income taxes.

I don't think I agree with your normative claim. I think ceteris paribus that it's a virtue of a tax system if people pay in proportion to the value they get out of society. It sounds like you think that virtue is controlling. I think there are other virtues as well.

People get value out of a society by consuming goods and services, income is how we as a society measure the value a person provides.

Income Tax taxes value provided, Consumption Tax taxes value consumed.

-- Consumption taxes can also be scaled/slabbed with the amount consumed, they don't have to be regressive.

Most of the benefits I've received of my higher income are intangible: social security, medical and dental security, mental well-being, confidence, social status, trust, financial fortitude, and respect.

Social security and medical security are tangible. I don't know what "financial fortitude" means.

The remainder are things money can't buy. A beautiful woman who is broke has most of them - should we impose the same taxes on her?

Taxing land is ideal. It's easy to collect, mostly held by already rich people, and there's zero economic deadweight loss.

(Actually, there's even an economic benefit to a land tax, as the tax encourages more efficient use. A _negative_ deadweight loss.)

The laundering they are after would overwhelm even BTC volume.

If someone brought BTC in China and told it in the US none of it would be reflected on the blockchain.

From the exchange's database's perspective, it'll be a credit to their CNY account and a debit to their USD account. No BTC needed to be moved.

Thus the BTC volume isn't a valid measure of the total value of transactions done in BTC.

can you elaborate? I'm not sure I follow

I'll number each individual step and you tell me which step isn't clear:

1. I wire 4000 CNY to the BTC exchange.

2. BTC exchange receives my CNY, and shows that I have 4000 CNY in my account.

3. I buy 1 BTC with my 4000 CNY; exchange shows that I have 1 BTC in my account.

4. Later on, I trade that 1 BTC for 600 USD; exchange shows that I have 600 USD in my account.

5. I transfer the 600 USD into my US bank account.

In step 3, since I never took possession of the BTC, it never leaves the exchange, and thus no blockchain transaction took place. If no blockchain transaction took place, then anyone observing the blockchain would be oblivious to my fiat to BTC trades.

Thanks. Step 3 and 4. You wire them the CNY, and how do they get your new BTC? I guess you're saying that the way most exchanges work is with a central depository of BTC and they just internalize (book out) their orders? (Unless the exchanges are really pseudo-derivatives dealers selling CFDs?) Isn't this exactly why places like Coinbase require your identity? (Does BTCe?) And if they aren't internalizing orders, particularly for massive sums which will be hard to find another side of the trade to internalize against, wouldn't that require a blockchain transaction to get more BTC in their depository, unless the exchange starts also acting as market maker, a la EnronOnline?

If I go to BTCe with a $50mm USD deposit, but they only have $25mm BTC equivalent deposited there, and "buy" an equivalent amount of BTC, will they not be required to obtain $25mmUSD equivalent more BTC from the blockchain?

Furthermore, do the exchanges take a position either long or short? If I then liquidate my $50mm worth of BTC, how do they get the fiat to wire to me, if they don't transact with somebody else?

I just fail to see how the major shifts in supply/demand are totally absorbed by the exchanges and never show up on the blockchain. Do the exchanges have an inter exchange market off of the blockchain or something? I agree with you, technically, that for small sums, there's no need for this and so the blockchain volume may not represent total size of the market, but at the scales we are talking about (see how much HSBC laundered for a single drug gang alone), how does this not show up, or rather, how does it never hit the blockchain? Particularly when somebody eventually wants to liquidate back to fiat?

"HSBC was accused of failing to monitor more than $670 billion in wire transfers and more than $9.4 billion in purchases of U.S. currency from HSBC Mexico, allowing for money laundering, prosecutors said."

That is one subset of the problem. BTC has what, $12mm/day volume? This one subset is over 2 years worth of all the BTC volume. Honest question, is off blockchain BTC trading really that big?

In the US, property is already taxed at the state and local level in most states. A Federal property tax is probably not constitutional.

what you describe is extremely unfair. my situation - scraped all cash I could earn in last 12 years of full time IT development/consulting, took a massive mortgage and bought a flat in my dream location in mountains where we go almost every weekend. I will be paying for this for next 25 years till its mine. It will cost me a fortune, it's quite risky move and if things will go very south (ie my health will deteriorate dramatically), I am more than screwed. This will be most probably the biggest investment in my whole life.

I live in rental apartment and probably will keep doing this (where I live and work I can't afford to buy, and won't for various reasons). By your logic, I should be taxed to hell just because you don't like seeing foreign capital buying off properties in London or some other location you want to buy property in? No, thank you. I've lived through communism, resemblance is striking.

I'm glad this is the topmost comment.

Punishing the 99.5% of good actors to sometimes discover the 0.5% of bad actors seems exactly like forcing people to prove their innocence as all are presumed guilty.

"Punishment" and "guilt" and "innocence" are irrelevant concepts here. Banks aren't being "punished" by being required to implement money laundering controls. It's a condition of doing business as a bank, to serve certain government objectives.

More generally: no western government recognizes a "fundamental right to commerce" such that restrictions on commerce require the existence of "bad" conduct worthy of punishment.[1] There is nothing "bad" about not implementing money laundering controls, just as there is nothing "bad" about talking a left turn on Constitution Ave between 9-5 on a weekday. But the government doesn't need to show its "bad" to implement the regulation.

[1] Though we kind of tried that in the U.S. under the concept of economic due process.

> More generally: no western government recognizes a "fundamental right to commerce" such that restrictions on commerce require the existence of "bad" conduct worthy of punishment.

That's very interesting. I'm not going to argue with your claim, because I assume you know more about it than I do. But it's a very strange idea, because in many ways commerce is more fundamental than anything in the Constitution or its amendments. Take away everything in the Bill of Rights, all the freedoms defined in the Constitution and our heritage of common law, take it all away and you can still have a society. It would be a terrible, awful society that no one would choose to live in (probably something like the DPRK) but it would still be a society. But without commerce, it is simply not possible to even have a society. I would go even further and say that a vanishingly small percentage of people would even be able to survive without commerce. So in a certain sense it's just as important as air, food, and water.

I don't think commerce is more fundamental. Humans survived for hundreds of thousands of years without commerce as we know it. Even at the time of the founding, the vast majority of Americans were living on mostly self-sufficient farms. Commerce in the modern sense is enabled by government: protection of property rights, imposition of a common currency, protection against pirates on the high seas.

You can argue that government isn't necessary to enable commerce. That you could have a world that has commerce without government. But even if that were true, in the world we do have, commerce is built upon the infrastructure provided by government. Just like the road system. And the government is thus entitled to regulate commerce just like it regulates left-hand turns on roads.

Of course, this is just an argument that regulation of monetary transactions is consistent with the assumptions underlying our institutions. I have no opinion on whether money-laundering controls are a good idea or not.

> Even at the time of the founding, the vast majority of Americans were living on mostly self-sufficient farms

People still bartered, which is recognized and taxed as commerce by the US government. They did not build or repair their own tools, for example - the village blacksmith was usually the most important man in town (or the region). You could argue that once they had a set of tools and land, people could be self-sufficient, but they still engaged in trade. And they would not have been able to be self-sufficient without trading beforehand - how else would you get a horse?

But if you ordered some horseshoes made and then refused to take delivery, the blacksmith would still sue you in a government court.

There are a wide variety of commerce systems that require no government courts. Consider the Silent Trade, the old system used by Africans and Indians to trade.

https://en.wikipedia.org/wiki/Silent_trade

Realistically, in Colonial America, reputation systems and the like were probably used far more than courts.

In fact, even today, extra-legal systems are widely used. My company buys SAAS services primarily from the US and Australia, but we have little ability to use the courts in either of those countries. How do we know Atlassian or Cloudflare won't take our rupees and run?

Back then, I highly doubt you would get sued. Your reputation would be ruined and you'd have trouble ordering anything else, but I think we're getting a bit away from the point I was trying to make.

"But even if that were true, in the world we do have, commerce is built upon the infrastructure provided by government. Just like the road system. And the government is thus entitled to regulate commerce just like it regulates left-hand turns on roads."

So by this argument, if the first amendment didn't exist, the government would have the right to regulate speech that uses public infrastructure (anything on public airwaves, the internet, delivered via the mail system, etc)? To make things more concrete, could the US government make a law "No one shall manufacture or sell a wallet"? (to pick the first thing that I saw).

Note that I personally am kind of ok with this, because the whole point of being sovereign is that you can do what you want; but I'm interested in exploring the justifications used to get people to buy into the system.

> > Back then, I highly doubt you would get sued

Common law contract and tort principles were well-developed back then, and indeed were created for situations like ordering horseshoes from the village blacksmith, or ordering a new axle for your windmill.[1]

> So by this argument, if the first amendment didn't exist, the government would have the right to regulate speech that uses public infrastructure (anything on public airwaves, the internet, delivered via the mail system, etc)?

No, because the first amendment does not create the right to free speech, but recognizes a right to free speech that already existed. But if that wasn't the case, then I think the answer would be yes.

[1] https://en.wikipedia.org/wiki/Hadley_v_Baxendale

Virtually all commerce in modern society is intermediated by fiat money, which is a product of the state. People are free to trade goods with Monopoly money if they want to, but if they want the security and backing of the United States, including the force and authority of its agents and courts, then they use the US dollar, and are subsequently subject to US law.

I'm not an expert on this topic, but I'm fairly sure that if you traded gold, or Monopoly money, or anything not fiat currency, for goods & services, you would still be taxed. The right to tax does not arise out of the use of fiat currency.

I'm not talking about tax. I'm referring to rayiner's initial point, which is that "right" and "wrong" have nothing to do with it. Businesses such as banks that intermediate financial transactions with state currency are subject to those states' regulations. It's not "wrong" to not enforce money laundering controls, but it is illegal, because you operate within US jurisdiction under a US charter trading US currency. If you don't want to follow the laws, go trade something else somewhere else. (To this point, the Russians and Chinese have been making a point for years of getting away from the dollar, specifically to get out from under the America's thumb.)

I agree.

Many honest people will not tolerate total financial surveillance and automated techniques like those in the article make honest people intentionally engage in structuring even though they have "nothing to hide"

Financial privacy is like clothing. Yes, clothing can be used to conceal contraband and further illicit activity that is almost universally regarded as bad, but that doesn't mean the entire world is willing to be strip searched.

So society would be better off without altruistic punishment in the financial sector, which by the way is one of the most effective weapons in the post-nuclear age?

I am afraid I disagree. Criminal activity has one clear shared goal - to make money. And banks are outrageously complicit in money laundering for criminal organisations.

Yet detecting crime through its laundering of proceeds is probably the most powerful Single idea we have - imagine a world where you could only commit crime in cash or for free.

We talk about a poverty free world. A crime free world is also possible. Just make sure it does not pay.

> A crime free world is also possible. Just make sure it does not pay.

Even fictional post-scarcity utopias have crimes of passion. You cannot get anywhere near crime free without one of the following options: a) absolute coercion and perfect preemptive law enforcement, b) no laws to speak off (including no prohibitions against murder), c) radical departures from human nature as we know it.

That said, I agree with your broader point: sometimes it is worth it to be able to detect and stop crime based on its financial footprint, even if you don't gain economically from doing so.

Completely agree. When I was working for a French bank, I saw how the US treats money laundering and it's an absolute nightmare. Let's take the example of a transaction that is processed in EUR by a French bank, from a French company to a Chinese company, on goods manufactured in, let's say, Ivory Coast. If in the email discussion, somebody answers while he/she in the US for holidays, then the transaction will fall under Fatca while there is absolutely no US interests affected whatsoever. Bottom line: Fatca is not only about money laundering, it is largely about maintaining the US dominance of international banking.

Fatca is one of the reasons American nationals can't open a bank account in my country of nationality (Belgium). Banks just can't be bothered to implement and pay the costs for this.

Financial institutions are not interested in being good at catching money laundering operations. They are interested in avoiding expensive fines. False positives are really bad for customer relations.

When a bank loses money they pay for it. It happens often, for millions, says a friend who works in fraud analysis.

You are talking about fraud (I'm assuming wire fraud, chargebacks, stolen accounts, etc.) which I would put in a different category of risk for a financial institution than money laundering. These activities would lead to direct losses, whereas money laundering would not.

I assume allowing laundered money through your institution is of less concern, since it would not lead to a direct loss to the institution. This is assuming you are not knowingly allowing the laundering to occur and you are following all AML and KYC regulations meant to prevent money laundering. If these standards are not met, it may lead to fines and having regulators watch your institution more closely.

Wire fraud and stolen account schemes generally get eliminated/caught because of laundering issues, not because of how they do the actual fraud; and for large cases of fraud / large stolen accounts, the laundering capacity limits the total amount.

Good AML provisions help reduce your fraud losses, even if you decide to make some "exceptions" for your VIP customers.

It's actually the opposite of fraud, because laundering money involves transactions which frequently make a great deal of money for the bank. Look at the fat fees Goldman Sachs earned setting up the 1MDB transactions for Malaysia. There was a lot of speculation around the financial crisis that banks opened their doors to money launderers because it was the only available liquid asset available in sufficient quantity to keep them solvent. So the bank actually has an incentive to launder money, so long as it can do so without being caught and fined or having its employees go to jail.

The point below is true though. Many of the same markers of fraud indicate money laundering, and having a good risk program can reduce both if that's your intention.

Fraud and money laundering fine avoidance are two different things.

Certainly the only moral thing to do is the absolute minimum to avoid being shut down. Money laundering is not ethically wrong and making it a crime is a great example of government overreach.

Money laundering is not a crime? It's required to finance certain crimes, or to cover them up.

It's not victimless.

It's like civil forfeiture. It's LE being lazy or saying they can't figure out the real crime, so they'll just go after people's assets instead.

Private communications are also required for certain crimes, or to cover them up. Should we require that Apple monitor phone conversations for evidence of crime?

Wow. Since when is every regulation supposed to pay for itself? Do police? SWAT?

I highly suggest reading Steve Randy Waldmann's analysis of the Baltimore riots and altruistic punishment. He is an actual economist, whereas this thread is as if he came in here and was telling you your code sucked.

For those curious, his blog is called interfluidity.

With everything there is a trade-off. For some the lack of dollars brought in (or what have you) outweighs the benefits.

An example is TSA. Some believe that even one life saved by the extra security (theater or not) is worth having to take your shoes off. Others do not.

I just followed your advice and read Steve Randy Waldmann's analysis of the Baltimore riots and altruistic punishment.

It is very clear that he is biased in his analysis. He is also wrong about the Baltimore riots being "altruistic punishment", at least according to the strict definition here[1], as the people rioting claimed to be directly harmed by the existing law enforcement structure that they were rioting against.

Finally, why do you think that simple economic concepts can not be understood by developers? Why do you think there there are not lots of people on here with business degrees or backgrounds in economics?

[1]https://en.wikipedia.org/wiki/Third-party_punishment

I think his point is that the idea that knowing you will be retaliated against, even if it's a "net loss" in some ways for those retaliating against you, is an extremely powerful force for good in our society. Do you disagree? I think that because I'm a quant and I see people making pretty basic mistakes here, or making fallacious assumptions about other fields assuming they follow the same logic as software, quite often, and that seemed to be happening in this thread. I maintain that developers are usually not qualified to talk about economics in much beyond the most basic sense, because they haven't actually studied advanced economics beyond how it pertains to the tech industry. They just make linear extrapolations and are shocked (shocked!) when their narrow (and admittedly unique) model of the world breaks down.

> I think his point is that the idea that knowing you will be retaliated against, even if it's a "net loss" in some ways for those retaliating against you, is an extremely powerful force for good in our society. Do you disagree?

If you have taken anything beyond economics in the most basic sense, then you know the answer to any economic question is, "it depends."

So in this case, I do not think that the riots did anything to the city government's or local police force's sensibilities, because it doesn't cost them anything...they get to use the house's (the citizens) money. It is the same reason lawsuit settlements do not seem to have an effect on police misconduct -- they don't care, it doesn't really cost them anything. So to your direct question, regarding the Baltimore riots, no I do not think it was a net good. The people protesting got nothing, businesses and private property got trashed, and the city paid costs with taxpayer's money, and nothing changed. There was nothing good that came of that, or the threat of that.

I do take your point about armchair economists seemingly having the solution to nuanced macroeconomic problems that stymie experts, and I agree we have seen a lot of that recently with Brexit.

However, I do not think that people can't have an opinion on something because their eduction level in that specific area is not elite enough, better to explain why they might not be right, if you have a better answer.

Maybe you could make them one of today's lucky 10,000.[1]

[1]https://xkcd.com/1053/

Hold on you really don't think the Baltimore police will back off or hesitate before they do what they did again? The only thing that influences their decisions are if it cost them money or not????

Of course for a short time they will, because they do not want to be fired for bringing unwanted attention.

But yes, I believe that will be short lived. There are no real consequences to their actions, and the riot did not change that.

So if you think a riot and damaging private property for a short window of kid gloves if you are black from the police is a win, then ok.

And to be more clear, yes, if penalties came out of their pensions instead of taxpayers pockets, they would change their behavior long-term.

You don't think having to stand down rioters, have molotovs and stones thrown at you by an entire community of people proclaiming they are fucking done with dealing with you around, have shots fired at you, have officers shot in their squad cars has any effect on long term behavior? Sorry I'm just sitting here scratching my head.

I can't tell if you are being serious or not, but there is no mystery and this should not be confusing to you.

Baltimore was far from the first "ethnic" riot[1] in this country and none of them have ever caused a lasting change in police behavior. If anything, they just reinforced the notion among police that militarization of police is necessary. And I stick by this track record for my analysis of the situation.

Now, there is some very recent activity to support your analysis. Namely, Baltimore has released a new "use of force policy" in reaction to the Baltimore riots.[2]

I am not immediately conceding here, because it is common for local governments to make token "changes" to placate the agitated after flareups, mostly to save their next election. Furthermore, as the article points out, this change has no teeth, meaning there is no external oversight, so these are all more "recommendations" than anything else. The attorney for the ACLU expresses the same skepticism that I have regarding the recommendations.

So will Baltimore be the first to cause lasting change in local police behavior? It remains to be seen, but if it does, it will be the first. Showing that historically, riots do nothing for long-term change.

[1] https://en.wikipedia.org/wiki/List_of_ethnic_riots#1978_to_t... [2] http://www.baltimoresun.com/news/maryland/crime/bs-md-ci-pol...

There seems to be some bizarre skepticism by some HNers of financial institutions' AML practices. But the thing is, AML is really difficult and expensive for banks, hiring floors of consultants dedicated to AML. This is a billion dollar problem (or more?). Whatever about training a CNN to classify images, an effective ML model/pipeline to tackle AML will break the back of a huge problem.

It may be a billion dollar problem but it's a billion dollar problem like tax software, it wouldn't exist without the troll demanding his toll. The resources could be used elsewhere like on curing disease or on educating people: http://www.americanbanker.com/bankthink/money-laundering-is-...

It's a billion dollar problem that doesn't need to exist.

Many people will probably make cursory argument that taxes(and by analogy ML regulation) are a necessary evil(They're now doing the troll's work for him)— I'd argue they are not necessary and they short-circuit us working on better replacements.

But in the real world AML isn't going away, and it continues to be expensive.

The troll is money laundering, not the regulation. And you can't ignore the troll, or even get rid of it, so you must mitigate damage.

How so?

"Money laundering" is an imaginary crime. If certain patterns of transactions were not arbitrarily outlawed, there would be no need to enforce this law.

All kinds of fraud and theft of money from credit card and banking account data are very real crimes, and for them the main effective method of detection is following the money trail, and the main effective method of prevention is making the laundering hard, since already laundering is the difficult part that limits the scale of such operations, not the actual theft/fraud.

Wonder why there is such thing as spam of "earn $$$ working from home handling bills online" ? It's because someone there has excess capacity to steal money, but not enough capacity to launder it.

Only in the sense that all crimes are "imaginary." Crimes are acts made illegal by law, which money laundering is. It's as illegal as murder.

Certainly you're not comparing tax evasion to murder?

Insofar as the logical basis for classifying them as "crimes," yes, I am. Not in terms of seriousness or morality, obviously. Also, conflating money laundering and tax evasion shows a serious lack of understanding. Tax evasion can involve money laundering, sure, but money laundering is often the mechanism used to enable far more serious crimes, including, yes, murder.

Many things "enable" murder, including cars, guns, bullets, knives, etc. In any case this utilitarianism is a dead end. If it is not absolutely fungible, it's not money, it's coupons, food stamps, tokens, but not money and will not ultimately be used as such.

I agree and would add that trade via currency is in all cases voluntary. Prosecution of "money laundering" is a standard tool in the tool belt of all collectivist regimes who ultimately find it necessary to attempt to control as much activity as is technologically feasible.

There is zero incentive to recognize money laundering "too well", particularly because it means essentially whatever FinCEN wants it to mean. It's an issue for lawyers to hash out exactly what patterns must be recognized.

There may not be an incentive to recognize money laundering "too well" but there is incentive to get better. FinCEN will continue to raise the stakes of what is an acceptable AML program and staying ahead of the curve reduces the probability of a future fine.

I can't find the reference right now but someone once wrote that PayPal was, at its core, one of the most sophisticated fraud detection engines ever created.

In their early days, they almost went under multiple times because of the millions in chargebacks (This is one of those impossible problems that Elon Musk helped solve).

There is no such thing as money laundering. There is such a thing as food stamp fraud.

Reply on news.ycombinator.com