Comment by mattgibson | Hacker News Reader

Comment by mattgibson | Hacker News Reader

Parent Full thread

> Poverty is the metric that should be focused on, not inequality

No. Inequality is the problem. Poverty has nowhere near the same damaging effect.

This book is essential reading on the topic:

http://www.amazon.com/Spirit-Level-Equality-Societies-Strong...

The authors use two test-beds: a comparison of countries around the world and a comparison of states within the US. With these, they show that with overall wealth held constant, it's the larger inequality that causes more social problems.

For example, the UK and Japan have roughly equal GDP and similar average incomes, but Japan is less unequal. On almost any measure you might choose of how well a society is doing, Japan is better. Mental health, drug use, teen pregnancy, violence, divorce, abuse, the whole lot. The same goes across all of the states and countries they looked at.

The sheer scale of the evidence is overwhelming. It's very rare in the social sciences for literally hundreds of studies to point to the same broad trend across so many different parts of the world.

Their work is a revolution as it overturns the common assumption which you expressed that poverty is the problem.

From an amazon review:

This book presents the thesis that many ills of today's society (obesity, mental illness, rates of drug abuse, ...) can be attributed to large income inequality. The authors make this point using two-dimensional scatter plots, with income inequality on the x-axis, the prevalence of some form of social ill on the y-axis, and dots in the plot representing individual countries. These plots generally show a positive correlation between income inequality and various social ills.

As a statistician, I would like to comment on the soundness such argumentation: unfortunately virtually all graphs are plagued be a confusion of correlation with causation.

Reviewer then explains correlation and causation. I could claim "inequality is good" using the same methods. I'd argue inequality is immaterial but does tend to be correlated with poverty, which is what we should focus on.

Not to be outright dismissive (I should read the book), but this is a problem in social sciences in general. I think software simulations would be a lot more useful at this point than scatterplots comparing two variables used by so many economists today. I think I've read about people starting to do that in various areas, not sure about how prevalent it is in economic modelling. AFAIK not very.

Yeah that's a fair criticism, although as someone who has spent 10 years teaching social science, I have to say that was at the forefront pf my mind when reading it.

I was still mightily impressed with their data.

The range of studies is extremely broad (into the hundreds), and there is a very wide range of experimental procedures in use. Not all are correlations.

>inequality is immaterial but does tend to be correlated with poverty

This point, they explicitly refute with a vast array of data. In cases where there is both poverty an inequality, there is a problem. When poverty alone, the problems are massively reduced.

I read the book with a critical eye. I've taught social science for the best part of 10 years and this was not lacking in statistical rigour. Other academics in the field who I've discussed it with agree.

I think you'll enjoy the book. It challenged a lot of my preconceptions to the point where I changed my mind, which is was unexpectedly refreshing.

Causation does not exist in science. Causation only exists in theories that reside in our head. If you want to rely on evidence, and you want the notion of causality, you're in trouble because you can't have both, see the entire history of western philosophy in particular David Hume.

> Causation does not exist in science.

This is so over-simplified it's wrong. If I drop a weight, what happens? Is there a causal chain there? Would an animal unable to perceive a causal chain be able to live very long?

Okay, show me an experiment that verifies our intuitive notion of causation. Cause is a philosophical principle that exists like I said only in our heads and it's a question about the future. If I do this, this will happen. Science is about evidence, experiments, and the past: we did this and this happened.

If you are going to be doing science, you will not prove causal links. You will develop a theory which explains some phenomenon and implies some predictive assertions and then provide evidence for that theory and test the assertions. Cause is much deeper than science can currently tread.

You've defined causation in a very weird way that excludes what you think science covers, and I don't know why you did that.

I ordered the book, but can you share a thumbnail of what they propose as the causal relationship between inequality and poor performance on societal indicators?

On first blush it seems too neat a solution. Just look at Texas—massively unequal, there are oil barons and illegals within the same zip code, but Texas is doing well because it's possible to live the life style of an upper middle class person in NY on a pair of civil servant salaries down south.

Sure.

Their hypothesis is that as social animals, we are wired to respond neurologically to status Higher/lower status individuals experience dramatically different levels of certain neurotransmitters, which is something that they cite animal studies to support. Experimentally manipulating the social environment e.g. by placing a monkey in cage where it was surrounded by either smaller or larger peers caused observable changes in brain chemistry as the monkey moved from alpha male status to subordinate and back again.

There's a good argument for an evolutionary role here, as you can see that being on the bottom of a social hierarchy may be dangerous, so those who experienced it as stressful and were motivated to escape it somehow may have had higher chances of survival.

Being trapped in such a position appears to cause a deeply stressful state of mind, where people are literally unable to function normally. This doesn't happen when people are more equal as the perceived status differences are small, so the effect is lessened.

> The sheer scale of the evidence is overwhelming.

Really?

Because your description sounds very flimsy. Can you really compare Japan and the UK without correcting for a bunch of confounding cultural variables that have nothing to do with equality?

Is there no society that is "more equal" than Japan, and yet scores terribly on the metrics you suggest?

Reply on news.ycombinator.com