Re: Federal Expenditures

First off - why did you avoid my question? Is a 0.3 better than a 0.4?

Second, what is this "social ills" thing you keep spouting.

And third, did you read the BMJ article by or do you normally really on
someone elses sound bite. Your links are to blogs, not the actual
articles.

I do not agree with the blogs interpretation that income distribution
is causal to health status.

For example, the blog makes this interpretation of Kaplan:

"Adjusting these results for average income in each state did not
change the picture: in other words, it is the gap between rich and
poor, and not the average income in each state, that best predicts the
death rate in each state.

First you need to appreciate the part-correlations left unreported but
that can be inferred by the data. Second, you need to be aware of the
log-median correction for income used by the authors has some
statistical issues. Next, when correcting for race, the authors err in
their method and in their interpretation of what the results show.
This issue also undermines the use of other covariates. And, most
concerning from an analysis perspective is given the plethora of data,
why they did not do a simple stepwise regression? Running pearsons is
inappropriate. I have a problem using pearsons with

I have no doubt that there is a correlation between income distribution
and outcomes - but I am not about to suggest that poorer outcomes
result BECAUSE of how incomes are distributed. Kaplan agrees with me,
not the blogster, when he says:

"While the present results do not prove that income inequality causes
poor health, the results are dramatic and suggestive enough to make
further research in this area a high priority."

So, what did Kenedy find in his study?

Well, he used pretty much the same data from fairly similar sources at
a similar time, and he states that:

"Gini coefficient showed little correlation with any of the mortality
outcomes in these data, with the exception of homicide."

He did find, using his home grown Robin Hood coefficient that there are
correlations. It makes one wonder why he created this new measure,
doesn't it?
..
He argues that it's because the GINI is sensitive to the tails of
distributions - which is one argument that almost, but not quite, makes
sense - and it speaks to the inherent lack of robustness of the
measure. You might want to look up the statistical meaning of the
term, by the way.

He makes an interesting statement, as well:

"Income distribution may be a proxy for other social indicators, such
as the degree of investment in human capital. "

I would argue that overall income is a much better indicator of
investment in human capital. Though both authors use median income as
a covariate and discount its effect on GINI or RH, neither uses total
per capita income as a proxy for this nebulous human capital investment
construct. The reason I find this relevant is in viewing the high
GINI/RH states and what I know about the economics of those states, to
me it makes for a plausible hypothesis.

Where your pundit shows us what he thinks, the data are not nearly as
convincing.

BTW, in the same BMJ volume there is an editorial by Smith and another
paper by Ben-Schlomo worth reading.

You may also want to read Kennedy's piece four years later published in
BMJ -
Where he says:

"A combination of the individual income and neo-material
interpretations is a better fit to the available evidence on income
inequality and health, is more comprehensive, and has greater potential
to inform interventions that advance public health and reduce
inequalities."

His airplane analogy is quite interesting.

I always prefer the primary data to someone elses spin, don't you?
\
So, is a .4 worse than a .3?

You never answered the question.

js

.

Relevant Pages

  • Re: Federal Expenditures
    ... I do not agree with the blogs interpretation that income distribution ... is causal to health status. ... I have no doubt that there is a correlation between income distribution ... "Gini coefficient showed little correlation with any of the mortality ...
    (soc.retirement)
  • Re: OT: Cleveeland cop dies on the line
    ... condition of increasingly unequal distribution of income exists. ... Commonly income distribution is expressed by a value called the ... GINI index or coefficient. ... such as crimes against property, crimes against persons, and ...
    (rec.crafts.metalworking)
  • Re: LVT dustup over at AngryBear
    ... >> I think the tax burden should be shifted towards the upper income end. ... > So, as you wisely point out, the real question is: ... > according to what the income distribution actually _is_. ...
    (sci.econ)
  • Re: LVT dustup over at AngryBear
    ... Income tax rates are usually associated with income brackets. ... income above $1B without noticeable effect on economic activity. ... So, as you wisely point out, the real question is: ... according to what the income distribution actually _is_. ...
    (sci.econ)
  • Re: Gap in Life Expectancy Widens for the Nation
    ... The central point in this program was that while higher income people do ... disparities in life expectancy for richer and poorer Americans, ... Life expectancy for the nation as a whole has increased, ... official statement of national health objectives issued in 2000, ...
    (soc.senior.issues)
Loading