Re: Federal Expenditures


Islander wrote:
js wrote:
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

You cite legitimate concerns about the direction this research has taken
and I'm sure that these concerns are reflected in the peer review of the
research (not the secondary reports of the research). As I said, this a
a tool in a field where there are not any absolutes. After all, this
line of research is not very old.

The issue is one of interpretation for political purposes. In most
cases, in my experience at least, the interpretation fo a correlation
or association becomes causal if it serves a political purpose. In
addition, this causalities are espoused in the absence of understanding
of the metric or its limitations even when authors are careful to
express those.

Most controversial, and certainly most challenged, is the correlation
between income inequality and health issues such as mortality rate.
Just because there is a correlation does not prove cause and effect as
Granville pointed out in his '98 paper.

But your bolgmeister does not take such care.

Other hypothesis might also
explain why we have a higher mortality rate in this country than that in
other countries having a lower income inequality.

I think inequality of income as measured by the GINI coefficient is a
biased measure. It has inherent problems in measurement and these have
been detailed elsewhere.

The relationship between income distribution and outcomes likely exists
but not because of the distribution but because of other factors that
arise in areas where such distributions exist.

For example, homogeniety of income may well reduce crime rates simply
because there is nothing of value to steal. Homegeniety of income may
improve educational outcomes because there is no tail. There are
plenty of alternative explanans for the effect.

Suggesting that social policy designed to reduce heterogeniety would
improve other factors, therefore, is inappropriate and likely to fail.

We in this ng have
certainly debated the cause and effect of the cost and effectiveness of
health care in this country compared with others offering universal
health care. While the critics of this research question cause and
effect, I'm not aware of equivalent work that identifies other cause and
effect with the same degree of rigor.

Mortality is associated with health care. There is no doubt about
that. However, in industrialized countries, the structure of health
care financing is not the sole or even primary determinent of mortality
rates. Those who argue that it is make the same mistake as you have
made assigning causality of your "social ills" on variance in an
economic inequality metric.

And as an aside, longevity is not all its cracked up to be. it can be
very expensive for society and this is probably the largest challenge
facing the industrialized nations today - how to afford a lot of old
people from a pension and health care prerspective - and this is
irrespective of social structure.

The arguments basically challenge
the conclusions without proof of some other cause. Perhaps you can
recommend citations?

I like to use the NHS approaches to treating cancer as an example
where fundementally different approaches to health care financing are
associated with glaring differences in outcome.

Also, as I said, the trends are more interesting than the absolute
thresholds cited in the article. Is a .4 worse than a .3? I would say
that a .4 trending toward a .5 is an indication that a problem exists.
Would you disagree?

The one thing that is more inappropriate than using the GINI as a
static measure and drawing conclusions is using it as a longitudinal
measure and inferring trends.

You recognize, I hope, that GINI looks at class membership at a single
point in time and as such says nothing about class mobility. In fact,
some authors have discussed that class mobility is much more important
than static membership. This argument has been applied well outside
the realm of the GINI - it is prevelant in the economics literature. I
recommend reading Herfindahl.

A GINI of 0,5 you cite as a problem, I would argue that such an
interpretation is inappropriate as it immediately infers causality.
You said that this was wrong, yet you do it yourself.

How about a 0.4 trending toward 0.3? Is this "good"?

Well, I am the King and this is how I do it. In my 0.4 society, in the
most simple terms, we know that the top 30 percent has about half the
income. This is unacceptable. I take the income of the bottom 30% and
give it all to the top 70% making the bottom 30% slaves. My GINI is
now 0.3 (again, I am not using the appropriate calculus, just a
simplification of the measure to make a point.)

OK - so it's outlandish, but I think I made my point.

I might also add that I appreciate your awareness of this work.
Definitely refreshing to find someone who has actually done some work
rather than simply voice opinions. This is not my field, but I find it
interesting (and I admittedly have opinions). Have you done work in
this field?

My roles in life have been varied but there was a point when I
consulted with various policy organizations regarding economics - some
NGO, some government, and quite a few private sector. Most was micro
econ in private sector healthcare.

Now they call every now and then to ask questions. My goal is to get
my handicap below 15.

My own interest was prompted by work on a project dealing with
affordable housing in our island community and similarities to
situations faced by other regions such as Aspen, the Bitterroot Valley,
Nantucket, Block Island, and other regions experiencing displacement of
their workforce by the high cost of housing. I am very much aware of
the "social ills" that can result when a region becomes economically
stratified.

I don't see this as an ill. In fact, I see this as an asset to
communities and countries - it demonstrates overtly the benefits of
success. You keep forgetting - I find rewarding success a good thing.
I think income distribution variance is a positive. I think that with
the ability to move across class lines, high GINIs stimulate the
economy and low GINI policies cripple growth.

I cited the work that has been done on health since it
appears to have attracted the most research, but there are also issues
of crime, drug abuse, corruption in the public sector, spousal and child
abuse, and other social problems. As I stated above, it is very
difficult to link cause and effect and if you can recommend serious work
that has been done in this area (pro or con), I would appreciate it.

Crime - effective deterent? Education and opportunity, not handouts
Drug abuse - considered by many as a disease (addictive behavior
disorder). Deterent? Treatment and education. This is a demand side
problem.
Corruption in the public sector? I think Randy Cunningham should have
been made an example of. This was a missed opportunity. I would
support term limits. I would support campaign finance reform.
Abuse - any abuse - is wrong. Criminal code enforcement.
Other social problems? I's be really concerned how we intend to pay
for Medicare and Social Security. This is the number 1 social problem
of the 21st century.

If I were King.....

js

.

Relevant Pages

  • 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: 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)
  • Re: more good news for U.S. economy
    ... median household income has fallen five ... better in 1975 and that our economy was in better shape? ... Job creation has not kept up with population growth, and the employment ... Rising health care costs are eroding families' already declining income. ...
    (rec.music.artists.springsteen)
  • Income and Education Are Bigger Determinants of Health Than Insurance
    ... Income and Education Are Bigger Determinants of Health Than Insurance, ... Report Shows ...
    (misc.health.alternative)
  • Re: OT: US elections - Was: Re: Pat Martinos minor substitutions
    ... To the extent that your employer pays for it. ... My take-home income ... Buying health insurance for a family of four would bankrupt me; ...
    (rec.music.makers.guitar.jazz)
Loading