[jorleif]

This is a very interesting idea and has some good potential for social change. So much energy, so misplaced. As a scandinavian my first thought is: Don't Americans realize that there might a point where it is cheaper to provide (or more exactly, force) better education and social security on those at risk, than increasing the police force and incarceration rates?

There is something that is not quite right with this narrative that Gladwell spins though. His past examples are about the winners of their era. On the other hand, the contemporary examples are losers who end up killed and squeezed by the police. Presumably, in the earlier mafia era there were also plenty of losers, we just don't hear about them.

The main thesis which I think is sound, though, is that innovation often (usually) involves a certain amount of breaking societal norms (which includes laws). It can take violent forms, such as the mafia, but also less violent law-breaking is often present. Think Napster, as a clear example but why not include Uber and AirBnB as well? Often innovation is not just about new technology, but rather re-negotiating social contracts. For this reason, when technological or social change introduces new economic opportunity, often it becomes populated by people who are willing to violate laws. I think as societies, we should try to figure out how to better utilize the violent and rebellious behvior of people, while limiting the detrimental effects. The same adrenaline addict might kill himself and a few pedestrian on a motorcycle when evading the police, but could instead be allowed to risk his life in some space exploration program or why not a technology startup.

[yummyfajitas]

Don't Americans realize that there might a point where it is cheaper to provide (or more exactly, force) better education and social security on those at risk, than increasing the police force and incarceration rates?

That was the theory pushed fairly successfully by many social reformers in the 60's and early 70's. The net result, or at least the concurrent event, was a massive crime wave.

The fact of the matter is that Americans are not Scandinavians. For homicides where the offender is known, more than half are committed by a demographic group that is pretty much nonexistent in Scandinavia.

http://www.fbi.gov/about-us/cjis/ucr/crime-in-the-u.s/2012/c...

Further, within the US, people tend to behave more similarly to where they come from (even if it was many generations back) than to some American average. I don't have data for any Scandinavian nations, but Tino Sanandaji has some data comparing Swedish Americans (a group which apparently self-identifies enough to be statistically significant) to Swedes:

http://super-economy.blogspot.com/2010/03/super-economy-in-o...

So it's very unclear that a Scandinavian approach would work here.

[dragonwriter]

> That was the theory pushed fairly successfully by many social reformers in the 60's and early 70's. The net result, or at least the concurrent event, was a massive crime wave.

It wasn't even a concurrent event, it was a preceding event -- the rising crime wave began roughly concurrent with the end of WWII, rose in a fairly smooth unbroken trend (With a few interruptions -- which started in the 1970s.) If any result in crime rates resulted from social reforms advocated in the 1960s and 1970s (which is questionable), its more like it was the interruptions in the long-running trend of increasing crime rates, not the increase that started more than a decade before the reforms were advocated.

Demographics in the key criminal age demographic -- both from the demobilizations after the WWII and Korea, and then the Baby Boom, is probably the key factor driving the increase (and subsequent decrease as that demographic bulge passed that age band.)

[vidarh]

> That was the theory pushed fairly successfully by many social reformers in the 60's and early 70's. The net result, or at least the concurrent event, was a massive crime wave.

The US have never invested in social security levels anywhere near the Scandinavian countries.

> more than half are committed by a demographic group that is pretty much nonexistent in Scandinavia.

Except that when you control for factors that correlate strongly with social status, the race effect disappears almost entirely. The US has a poverty problem first and foremost.

[yummyfajitas]

The US have never invested in social security levels anywhere near the Scandinavian countries.

How much spending do you believe is necessary to control crime via social security/etc? Is there some consumption level at which crime is expected to vanish?

Note that the US currently spends 60% of it's budget on redistribution and 4% on police protection. What would the optimal spending levels be?

http://www.usgovernmentspending.com/piechart_2012_US_total

Except that when you control for factors that correlate strongly with social status, the race effect disappears almost entirely.

What factors? Most of the obvious ones that I can think of (poverty, government dependence, unemployment) fail for this purpose since blacks do not make up anything close to 50% of people suffering from them.

Could you please state concretely what these factors are and why you believe the effect vanishes?

[vidarh]

> How much spending do you believe is necessary to control crime via social security/etc?

Enough to mostly eradicate poverty. I don't know what the cost of that would be in the US, but as long as you have widespread poverty I don't see a reason to assume you have tested the effect this would have on poverty. At least nothing like Scandinavian levels.

The US presently have a poverty rate about 3 times that of Norway (about 15% vs about 4.5%), despite Norway putting the poverty "bar" much higher - in Norway you are officially considered poor if your income is below 50% of the median. In Oslo that currently means about $24k/year. The US official poverty line for a single person in the continental states is $11,670 (which isn't that far from 50% of the median in the US either, but the cost of living in Norway is not twice that of the US), rising to $23,850 for a family of 4. With directly comparable numbers, the difference would be much greater.

Then you have to expect to wait at least a generation for reduced poverty to filter through to increased education levels, even assuming you fund the education system well enough that anyone who puts in the work can complete university for free.

Then you can start to get an idea whether or not Scandinavian level social security and education would make a difference in the US.

> Note that the US currently spends 60% of it's budget on redistribution

The problem is that you've created a system that causes so large differences in the first place. 60% to redress that is little more than window dressing.

Further, that 60% number is nonsense. A large proportion of that 60% - whichever slices you've added up to get at it - provides benefits for at least parts of those that pay into it in the first place, so the actual amount that is net redistribution is far smaller.

> What factors? Most of the obvious ones that I can think of (poverty, government dependence, unemployment) fail for this purpose since blacks do not make up anything close to 50% of people suffering from them.

Poverty rate certainly does account for a substantial part of it, with African-Americans being substantially more likely to be poor than non-hispanic whites, at ca. 27.2% and <12% respectively.

http://www.irp.wisc.edu/faqs/faq3.htm

Tack on education levels, unemployment, family cohesion and you get most of the way there.

[yummyfajitas]

Until you come up with a consistent definition of poverty, your theory is not even wrong. Recall that a theory is "not even wrong" if it's proponents can't explain a concrete set of steps to falsify it.

Now if you take an absolute definition, you need to explain why a lot of countries much poorer than the US (much of Europe) have less crime. Or why Brazil, Mexico and South Africa have far more crime than India. That's a tricky sell.

If you want to use a relative definition of poverty (e.g. what Norway does), then you need to postulate that the existence of someone earning more than you do within a national boundary is what causes you to murder people. I.e., the existence of someone with a platinum toilet drives people with gold toilets into a homicidal rage.

Poverty rate certainly does account for a substantial part of it, with African-Americans being substantially more likely to be poor than non-hispanic whites, at ca. 27.2% and <12% respectively.

Whites make up about 72% of the population, blacks 12%. Assume poverty is the sole cause of homicide, then multiplication suggests 3.25% of America is poor && black while 8.6% is poor && white. In that case, the black murderer:white murderer ratio should be 1:3. It's 1:1.

I don't know what you mean by "tack on". Do you have a coherent theory, or are you just hoping?

[vidarh]

> Now if you take an absolute definition, you need to explain why a lot of countries much poorer than the US (much of Europe) have less crime.

No, I don't, because I'm not postulating that poverty is the only reason for crime, merely that it is a substantial factor.

Another obvious factor is difference in legal systems that makes general crime rates extremely hard to compare (consider that the US has the highest percentage of its population in prison in the world)

Further, you're being imprecise. Many countries that are considered rich have high poverty rates. Such as the US.

> Or why Brazil, Mexico and South Africa have far more crime than India.

Apart from what I wrote above, all of these countries have massive poverty rates. I haven't checked all three, but South Africa's poverty level is at least under some measures at similar levels to India. Mexico also have a massive US-fueled drug war that accounts for a substantial proportion of all crime to the extent that it mostly swamps out most other factors.

But of course, this alone is not very relevant unless you - unlike me - assume poverty is the only factor.

> If you want to use a relative definition of poverty (e.g. what Norway does), then you need to postulate that the existence of someone earning more than you do within a national boundary is what causes you to murder people.

No, I don't. You're making up strawmen again.

The reason for pointing out the relative definition was to make it clear that the US definition (which is also relative - the specific numbers are adjusted regularly) and Norwegian definitions are not directly comparable, and that Norway's 4.5% number is vastly higher than it would have been under a US definition.

As such, we can not with confidence say that if the US brought poverty down to 4.5% after the US definition, even all else being equal, it would provide sufficiently stable living conditions to make it possible to reap whatever level of benefit the Norwegian system does from reduced poverty.

This was to address your issue of how much social security would be enough, after the ludicrous claim that the US tried to provide social security in the 60's and 70's, and that it didn't have any positive effect on crime rates, and my counter-claim that the US have never seriously tried to provide proper social security.

Also, despite your aggressive and rude way of asking for me to support my claims, you've provided nothing but that in defence of your claims. I take your aggression and rudeness and lack of support for your own claims as a good indicator that you have nothing.

> Assume poverty is the sole cause of homicide

Nice strawman. Pretty much your entire line of reasoning appears to be founded on setting up strawmen. Has anyone suggested poverty is the sole cause of homicide? No. I suggested that differences in poverty levels was one major factor confounding the claimed link between homicide and race. The rest is your own invention.

> I don't know what you mean by "tack on". Do you have a coherent theory, or are you just hoping?

By "tack on" I meant that, unlike what you seem to believe, I have never claimed poverty to the be the only factor. In fact, I originally pointed to social status specifically of those other factors.

I don't see the point in continuing this discussion and trying to explain anything to you, as you seem intent on misinterpreting every word, or you would have seen that "my" two theories are quite simple:

1) Poverty level (your own, and that of your immediate community, though they are usually largely the same) has a substantial effect on crime rate. You can falsify this theory by correcting for poverty levels in crime rate data, and see what difference it makes to the crime rate in the population reviewed.

2) If you adjust for factors that influence social status, the vast majority in the gap between crime rates for African Americans and white Americans will disappear. These factors include poverty, but also other factors such as education level and family cohesion.

The main point is not the specific set of factors, but that there are confounding factor that needs to be corrected for, that has dramatic effect on levelling the inter-racial differences in crime rates.

To falsify this, I don't need for there to be evidence of the specific effects of poverty on crime, as controlling for poverty and other such factors when looking at crime rates broken down by ethnic groups will either yield a result or not.

[wyager]

>Enough to mostly eradicate poverty.

What does that mean? A homeless person today can afford food, often (temporary) shelter, frequently a cell phone...

I'd say poverty has effectively been "mostly eradicated" already by production technology.

[IkmoIkmo]

What's your definition of poverty? It's important to be specific in your wording. For example there's absolute poverty (which seems to be what you describe, whether someone has the bare minimum to survive in terms of food, shelter) and relative poverty (which I think everyone can say, there's lots of that.)

But even on the first point, I don't think we've really eradicated poverty at all. There are millions of food-challenged people in the US, millions of people without insurance, millions of homeless. To say these people do not live in poverty because not all of them literally die of starvation is extremely myopic in my opinion. And that's in one of the richest and the most powerful country on the planet.

[Retric]

Federal budget is a long way from total government spending.

Further the vast majority of that goes to the elderly not the poor. Subsidizing the medical costs for a retired person living on 200k/year is hardly the type of social insurance suggested.

[yummyfajitas]

The link covers all govt spending.

[Retric]

And your link does not support the 60% number for redistribution. For example, you might think pension means SS, but states spend 200B/year on pension bennifits which are simply deferred compensation. When you read medical expenditure you might think Medicare and Medicare but local, state, and federal worker bennifits are limped there as is R&D.

Edit: I assumed you got the 60% from the federal numbers aka 24 SS, 22 Medicare Medicare etc, Safety Net 12% = ~58%. http://www.cbpp.org/cms/?fa=view&id=1258

[rmxt]

> For homicides where the offender is known, more than half are committed by a demographic group that is pretty much nonexistent in Scandinavia.

> Further, within the US, people tend to behave more similarly to where they come from (even if it was many generations back)

Is your implication really that black/brown people everywhere are (or would be) inherently more murderous? Perhaps I am incorrectly reading between your lines.

[yummyfajitas]

I don't know what you mean by "inherently". I do believe many traits are strongly demographically linked and that demographics tend to drown out other factors. I don't know the particular cause beyond that - I've seen very little convincing evidence one way or the other.

[rmxt]

If by "linked" you mean correlation, then sure, I agree there are links between demographics and crime/poverty in the USA. However, your implication seems to be that if Scandinavian countries had substantial black/brown populations, then they too would have murder rates comparable to America. That was what I was attempting to get at by "inherently". How does your conclusion follow?

[yummyfajitas]

My specific claim: some crime causing factors live within subpopulations. If country A has a high crime subpopulation and country B does not, it's a little silly to point to a policy in country B and blame the high crime in country A on the lack of that policy.

I'm simply rejecting the idea that differences in crime rates between the US and Scandinavia are due solely to policy. Demographics matter too. A little googling suggests there are wild disparities in crime rates based on demographics even in Scandinavia.

http://en.wikipedia.org/wiki/Immigration_and_crime

[IkmoIkmo]

What do you mean by demographics? If you mean age, sure, 10 year olds or 60 year olds show lower crimerates than 25 year olds. But you linked specifically to an article that discusses various ethnic migrants. So again, you seem to be saying that ethnicity matters in crime.

And to this point, we all agree, there is a correlation. But what most of us don't agree with, but it almost seems as if you're implying it, is that there's something inherent about the ethnicity that makes one prone to crime. As if there's a genetic predisposition to theft, violence, rape, murder.

So you've left at least me curious as to what you're really saying.

[ebfe]

The average IQ among black Americans is 85, and the average testosterone level among black males is 20% higher than it is among whites. Both of these traits are strong predictors of violent behavior.

[dragonwriter]

Even assuming those numbers are correct, measured IQ, at least, has pretty strong demonstrated influences from environmental factors (and many of the demonstrated negative influences from environmental factors are from factors that are products of poverty). Not sure about testosterone levels, but at least the IQ number is not really contrary to the idea that the underlying problem is linked to poverty and social disadvantage, not inherent to the population at issue.

[kamau]

>The average IQ among black Americans is 85, and the average testosterone level among black males is 20% higher than it is among whites. Both of these traits are strong predictors of violent behavior.<

The racists have come out to play.

[peterfirefly]

Hopefully you agree that it is an empirical question?

If it isn't possible to go out and look at the world to see if the average IQ is 85 then it isn't possible to see that it isn't 85.

Now, I wonder if somebody has done that and what the results were...

[kamau]

Spare me. Statements such as those are almost always concerned with promoting and upholding white supremacy/racism. Would the statement been given any legitimacy if it was stated that '...the average white IQ is 85...'? I doubt it. The fact that a statements like these are made with no or dubious evidence is bad enough. Asserting that there is an empirical question here, when said dubious statement is refuted, is the icing on the cake.

P.S. I do understand that this is an empirical statement/question. But so is the assertion that there is a teapot on the moon.

[vidarh]

Without bothering to question your numbers, this is entirely meaningless as an explanation for anything without correcting for education levels, poverty etc., both in terms of crime itself and in terms of effects on IQ and testosterone levels.

E.g. physical activity has an effect on testosterone, thus you might expect to see increased average levels of testosterone in a population that is more likely to have manual jobs, as you would expect in a population that on average is poorer. This effect may very well be drowned out by other things, but it is but one in a huge list of potential confounding factors.

[ebfe]

You could restate my viewpoint as the claim that poverty and education levels are entirely meaningless as an explanation for anything without correcting for hereditary intelligence, testosterone levels, etc., both in terms of crime itself and in terms of effects on poverty and education levels.

[vidarh]

No, they are not entirely meaningless, as you can demonstrate clearly the substantial effect they have by correcting for poverty, education etc. in the crime number and comparing the uncorrected / corrected results to see the effect they have.

When you then further break that data down by race, the purported intelligence differences and testosterone differences you claim either have no effect, are non-existing, or they have fallen away as a side effect of removing the other confounding factors, as eradicating socio-economic factors mostly erases racial differences in crime rate.

If they are real, have an effect, but fall away when correcting for other factors, then that suggests that these effects themselves are caused by the same confounding factors, and are not down to ethnicity after all.

E.g. it'd not be surprising at all if IQ test results in a poor, uneducated subset of a population where parents have less resources to invest in their childrens wellbeing, and where the chance of malnutrition is higher, are worse than in wealthier parts of a population. But the obvious place to look for causes are in the many known causes of low performance of IQ tests, several of which co-occur fairly frequently with poverty.

Any study that tries to pin such a difference on ethnicity is outright trash if it has not controlled very carefully for a rather extensive set of other potential factors. Including directly controlling for factors such as nutritional differences.

Another part of the reason for the outrage at claims like this, is that trying to pin down "black Americans" as an ethnic group is hopelessly flawed from the outset, given that "black Americans" genetically range from mostly European descent to entirely African descent, but with the vast majority being substantially mixed, and that discourse on that in the US is largely dictated by a combination of racism and cultural factors that one one hand makes it impossible for most mixed people to "pass as white" even if they have far more European than African ancestry, and on the other hand makes it culturally hard for a mixed person to identify as white without being considered a sell-out by the black community.

E.g. Obama is no more black than he is white, yet he is automatically considered black in the US by pretty much everyone, and would face flack from both "sides" if he were to try to describe himself as white.

Because racial designation is mostly a social construct that is largely down to peoples immediate knowledge of their ancestry, skin colour and social considerations rather than genetics (I'm reminded of the difference with the Portuguese, who during their colonial era and a higher willingness to openly intermix with the local populations had an intricate system of something like 12 different distinctions of levels between white and black that affected your social standing, which anyone who knows people with mixed children knows would be largely "luck of the draw"), it is highly geared towards being easy to manipulate (don't get the numbers you like? redefine "black" to find a definition that moves the delineation between groups in your study, and odds are you can find some dividing line that fits your agenda) which in itself makes any numbers suspect.

But even with a researcher without an agenda, without extensive genetic testing, result would at best tell you that people who self report as black fall in one group, and everyone else fall in another. Of course then you don't know if - assuming you had in fact isolated a trait that actually has a genetic component - the trait you are reporting on is actually a result of a genetic trait tied to "being black", or if it stems from the European intermixing into the "black" population that just isn't significant enough to affect the overall white population the same way.

I've yet to see anyone come up with research that claims such differences that does not fall flat on its face either by failing to address sufficient confounding factors, or by failing to at the very least discuss the considerations of how it ethnically delineated the population and why.

[tehwebguy]

I'm pretty sure that Scandinavian countries have males

[woah]

What crime wave?

[happyscrappy]

The Scandinavian approach could work in the US if the citizens were all of the type that would not cross an empty street against the light.

[vasilipupkin]

Note that just because something works in a country of 10mln, that doesn't mean it will work just as well in a country of 300mln. It's unlikely that Scandinavian style policies scale linearly. You can always find a sample of 10mln Americans that does just as well on all metrics as Sweden or better.

[TeMPOraL]

Last time I checked, the US had 50 states, giving 6 mln people per state on average. I think such solutions implemented on state level should work exactly as well as for a country of 10mln.

[vasilipupkin]

maybe. We just don't know. To assume that cultural factors have zero influence on success of such policies is probably incorrect.

[vasilipupkin]

haha. Why did I get downvoted for this? I am not implying anything racial or anything like that. I am talking about culture specifically. Illinois has a political culture of corruption that is probably worse than Sweden, as an example.

[k__]

Probably because most people think there are too many prisoners in the US and everyone who implies it should stay that way is seen dense/backward.

Didn't downvote, but I have the same opinion as those people.

[vasilipupkin]

I agree with you but that has little to do with swedish style welfare. Too many prisoners is due to draconian drug policies which are thankfully changing

[nawitus]

> It's unlikely that Scandinavian style policies scale linearly.

I think it's likely. In fact, by default policies should scale linearly from 10 millions upwards.

[slyn]

This is pretty reductive and ignores the fact that America is exceptionally racially, culturally, and economically different from region to region, state to state, city to city, et al.

Policy that works great in NYC isn't guaranteed to work great in LA. Laws that work well in Dallas or San Antonio might have negative effects in Detroit, Chicago, or St. Louis.

[IkmoIkmo]

By default? What default are you talking about? Sounds like fairy-tale default. Saying policies can just scale linearly from 17m people in the Netherlands to 1300m in China, needs an explanation a little more elaborate than 'it's the default'.

[nawitus]

Policies scale because resources scale linearly to the size of the population.

[vidarh]

Policies scale by compartmentalisation. A population of 1300m is not administered as one big lump of people. Nor is a 17m one.

[IkmoIkmo]

The point is, when you compartmentalize 10m people in 10x 1m provinces with local governments, it's relatively easy to have these 10 provinces be directed by a singular governmental policy. But when you try the same for a 100m population, or a 1000m population, you'd get 10 compartments of 100m, 100 compartments of 10m below this, and 1000 compartments of 1m. To expect each to follow the national policy is very difficult to do. You either need a strong dictatorship, or you need to accept that different compartments may elect radically different policies, which is why I don't think you can just implement and scale a policy to 1300m as easy as one can to 17m.

That's why you see huge differences for example between states even in things we expect to all agree on. There's very few national policies that simply scale to all states exactly the same, the military probably being a big one. But things like education, police or fire departments, housing, healthcare or even marriage, can be radically different. But here in the Netherlands? On tons of topics like education, police, healthcare or marriage, it's exactly the same in every single province, because we're pretty much one big state, despite quite large differences between provinces. e.g. 2 hours north-east, I can't understand the local language, here it's mostly protestant but two hours south is mostly catholic. But despite big differences in culture, language, religion, but also industries and demographics, it's extremely homogeneous in policies. To me that says that you can't just scale a policy to 300m like you can to 17m.

[gretful]

unfortunately for us, human behaviour doesn't scale linearly.

[fdsary]

Why don't the Americans do it on state or city level then?

Sincerely Swede

[vasilipupkin]

because of a complex interplay between Federal and state welfare programs. But also, they do more in some states and less in others. What I am saying is, if every state went to a Swedish model, it's not necessarily true that we would be better off. Paradoxically, the Swedes themselves could become worse off. Americans consume more because they have more disposable income, which means they can import things from Europe. What will happen if taxes are so high that Americans start consuming a lot less ?

[themartorana]

A huge boom in the middle class, as witnessed by the end of the Great Depression through the 1970s, during which time tax rates were incredibly high on high income as compared to today.

Also closing gaps in income inequality, and continued spending by that larger percentage of middle-class population.

Or so history teaches us, which we're rather happy to ignore now.

[refurb]

I would argue the economy boomed during that period in spite of the high taxes, not because of them.

[themartorana]

Absent any evidence, I could argue that's an absurd position to take, based on a desire for something to be the exact opposite of the truth, just to fill a flawed political bent.

[vasilipupkin]

absolutely, you are right that US tax rates on the very highest incomes were high. But overall tax burden was never as high. Remember, Sweden has both higher income taxes on the middle class than US ever had and high consumption taxes.

[vidarh]

Taxes in the US are not nearly as much lower than in Scandinavia as people like to think, though of course it varies quite a bit depending on state income taxes. When you take into account "necessary but privatised " bits like health insurance, it erases most of the remaining difference.

[true_religion]

The thing about health care is that it doesn't feel necessary when you're young, and healthy. Its when you're old, or are having children that you start to care about the health care system.... and by then its too late to revolt.

[vasilipupkin]

I don't think that's accurate, Swedish overall tax burden is something like 45% of GDP

[seanflyon]

For comparison US taxes are 24-27% of GDP according to Wikipedia.

http://en.wikipedia.org/wiki/List_of_countries_by_tax_revenu...

[jorleif]

I don't disagree, but experimentation should still be on the table. Maybe a certain state or municipality should be sponsored to try policies more along the scandinavian kind. Of course this is problematic. How do you for example stop everyone from moving into that particular priviliged district, etc?

[vasilipupkin]

states do have very different welfare policies. Some more generous than others. Texas has comparatively little relative to Massachusetts. Guess which way the net flow of people that are supposed to benefit from those policies happens

[MagicWishMonkey]

Poor people don't have the resources to move around the country.

Wealthy people are the ones who relocate and they don't benefit much from social welfare programs (if you ignore tangential benefits like not getting your car broken into by someone who needs money to eat).