# Wealth redistribution discussion (split from cartoon thread)

#

**1****Wealth redistribution discussion (split from cartoon thread)**

<i>Mod note: Original cartoon was posted here http://forum.dvdtalk.com/10095538-post374.html the discussion was split out to this thread to prevent the cartoon thread from getting bogged down

nemein

</i>

Lucky Ducky!

Apropos of that cartoon, here are two graphs that will show just how much we redistribute wealth in this country:

The Gini Coefficient is a measure of wealth distribution in a population. If the coefficient is 0, everybody has the same amount of wealth. If it's 1, it means that 1 person has all of the wealth and everybody else has none.

nemein

</i>

Lucky Ducky!

Apropos of that cartoon, here are two graphs that will show just how much we redistribute wealth in this country:

**Spoiler:**

The Gini Coefficient is a measure of wealth distribution in a population. If the coefficient is 0, everybody has the same amount of wealth. If it's 1, it means that 1 person has all of the wealth and everybody else has none.

*Last edited by nemein; 04-09-10 at 01:32 PM.*

#

**3****Re: The Random Political Cartoon Thread**

Now let's see how that graph correlates with GDP.

That's the best graph I could find on the internet. Maybe wendersfan will take pity on us and post something.

#

**4****Re: The Random Political Cartoon Thread**

And there does look to be a correlation in the graph, but I wouldn't know how strong. It seems obvious from just a glance that once you go over the 40 mark for the Gini index, you have only a single point at $15,000 per capita GDP, while most are bottom feeders. However, before the Gini 40 mark it appears that approximately 50% are at or above the $20,000 in GDP per capita. The graph looks relatively close to a "\" across the Gini score to me. Do you not see that? In fact, it looks like a fairly strong correlation. Go over 40 on the Gini index, and you have a per capital GDP and the best you can hope for is roughly the median of those below it.

#

**5**
#

**6****Re: The Random Political Cartoon Thread**

I'm not sure there is a relationship so much as there is a rule of thumb -- relatively unequal countries are poor -- that we manage to break. But you have low-Gini rich countries and low-Gini poor countries, which suggests to me that whether you're rich or poor has little to do with your Gini score. And you have low-Gini poor countries and high-Gini poor countries, which suggests to me that Giini isn't necessarily a function of GDP. But this is why we need wendersfan to run a regression for us.

*Last edited by JasonF; 04-09-10 at 09:55 AM.*

#

**8****Re: The Random Political Cartoon Thread**

Probably this afternoon?

#

**9**
#

**11****Re: The Random Political Cartoon Thread**

OK, I have the data in R. What do you want done with it?

BTW, I got the data from www.nationmaster.com, so it probably isn't exactly the same as the data used in Jason's graphs.

I'm going to lunch now, so I can do your analysis when I get back.

BTW, I got the data from www.nationmaster.com, so it probably isn't exactly the same as the data used in Jason's graphs.

I'm going to lunch now, so I can do your analysis when I get back.

#

**13**
#

**14****Re: The Random Political Cartoon Thread**

BTW, I got the data from www.nationmaster.com, so it probably isn't exactly the same as the data used in Jason's graphs.

I'm going to lunch now, so I can do your analysis when I get back.

#

**17**
#

**18**DVD Talk Hero

Join Date: Nov 1999

Posts: 36,981

**Re: The Random Political Cartoon Thread**

#

**19****Re: Wealth redistribution discussion (split from cartoon thread)**

<i>Mod note: Original cartoon was posted here http://forum.dvdtalk.com/10095538-post374.html the discussion was split out to this thread to prevent the cartoon thread from getting bogged down

nemein

</i>

nemein

</i>

#

**20****Re: The Random Political Cartoon Thread**

#

**23****Re: Wealth redistribution discussion (split from cartoon thread)**

OK, is this what you guys wanted?:

The correlation between the Gini coefficient and GDP/capita is -0.416 (n = 130.)

The correlation between the Gini coefficient and GDP_PPP/capita is -0.411 (n = 130.)

The correlation between the Gini coefficient and GDP/capita is -0.416 (n = 130.)

The correlation between the Gini coefficient and GDP_PPP/capita is -0.411 (n = 130.)

#

**24****Re: Wealth redistribution discussion (split from cartoon thread)**

To help give us perspective, here is an example of of a very strong relationship from Wiki.

<img src=http://upload.wikimedia.org/wikipedia/commons/thumb/e/ec/Anscombe%27s_quartet_3.svg/325px-Anscombe%27s_quartet_3.svg.png>

All 4 actually show a .816 coefficient.

So Wenders, tell us about the strength or lack of strength with a coefficient of around -.41

<img src=http://upload.wikimedia.org/wikipedia/commons/thumb/e/ec/Anscombe%27s_quartet_3.svg/325px-Anscombe%27s_quartet_3.svg.png>

All 4 actually show a .816 coefficient.

So Wenders, tell us about the strength or lack of strength with a coefficient of around -.41

#

**25****Re: Wealth redistribution discussion (split from cartoon thread)**

It's a strong relationship, e.g:

Two things:

1. All those funny dots and stuff mean that the hypothesis is assumed to be sound at the 0.001 level, meaning that there's a 1 in 1000 chance that it's spurious. It's a strong relationship.

2. But it doesn't appear to be a

Code:

> summary(lm.gdp) Call: lm(formula = gdp ~ gini) Residuals: Min 1Q Median 3Q Max -17158 -9618 -3841 4963 72752 Coefficients: Estimate Std. Error t value Pr(>|t|) (Intercept) 37440.4 5413.7 6.916 1.98e-10 *** gini -670.6 129.3 -5.188 8.11e-07 *** --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 14590 on 128 degrees of freedom Multiple R-squared: 0.1737, Adjusted R-squared: 0.1673 F-statistic: 26.91 on 1 and 128 DF, p-value: 8.105e-07 ------------------------------------------------ > summary(lm.gdp_ppp) Call: lm(formula = gdp_ppp ~ gini) Residuals: Min 1Q Median 3Q Max -18626 -9416 -2409 6800 60542 Coefficients: Estimate Std. Error t value Pr(>|t|) (Intercept) 37521.5 4976.2 7.54 7.60e-12 *** gini -606.0 118.8 -5.10 1.19e-06 *** --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 13410 on 128 degrees of freedom Multiple R-squared: 0.1689, Adjusted R-squared: 0.1624 F-statistic: 26.01 on 1 and 128 DF, p-value: 1.193e-06

1. All those funny dots and stuff mean that the hypothesis is assumed to be sound at the 0.001 level, meaning that there's a 1 in 1000 chance that it's spurious. It's a strong relationship.

2. But it doesn't appear to be a

*linear*relationship. What appears to happening is that at low levels of GDP/capita, there's no relationship at all, then it shifts up rather dramatically.*However*, I did a logarithmic transformation of GDP/capita, and then the relationship*weakened*. So I'm a bit baffled.