« Robust discount rates | Main | The correct metaphysical views about everything »
Sickness makes you poor
“I find almost no role of financial anything in the onset of disease,” Dr. Smith says. “That’s an almost throw-you-out-of-the-room thing,” he confesses, but the data, he and other economists insist, is consistent.
Income, says Dr. Preston, “is so heavily influenced by health itself.”
Here is more. The main point of the article is that education is strongly correlated with better health outcomes, although the author too quickly assumes a causal connection. For instance education may signal rather than cause low time preference and thus responsible behavior. In a slightly different health context, here is Jane Galt on causation vs. correlation. Is it worse to be overweight, or not to exercise much?
Posted by Tyler Cowen on January 3, 2007 at 07:01 AM in Medicine | Permalink
Comments
Perhaps the most interesting thing in that Campos article Jane Galt links to is the quite plausible implication that the greatest contributor to mortality among overweight people may be their propensity to diet. The tendency of the medical establishment to completely fail to distinguish cause from correlation when analyzing epidemiological studies is staggering.
Posted by: Dave H at Jan 3, 2007 8:47:45 AM
The tendency of the medical establishment to completely fail to distinguish cause from correlation when analyzing epidemiological studies is staggering.
Indeed. Though people on all sides seize on that. Several of the "fat doesn't matter for death rates" studies seem to ignore-- or at least have trouble taking into account-- that sick people frequently start losing lots of weight years before passing away from their illnesses. This of course moves obsese people to overweight, overweight to normal, and normal people to dangerously overweight. The fat acceptance folks seem to think that the studies reveal that "being underweight makes you sick" rather than "being sick makes you lose weight."
I certainly know people who were overweight and in poor health most of their lives, but lost quite a bit of weight in the few years before passing on, as they struggled for quite a while. It takes a careful study indeed to tease out the relationships.
Posted by: John Thacker at Jan 3, 2007 9:33:38 AM
This is an academic interest of mine, and I think it is indeed likely that more years of formal education really do *cause* an improvement of health and an increase in lifespan. www.hedweb.com/bgcharlton/ed-expansion (Although I would freely concede that this is not yet _proven_.)
The way it works, I believe, is that the core effect of formal education is to inculcate abstract systematic modes of thinking. These are not spontaneous for humans - we are naturally animistic, and see the world mainly as a web of social relations.
On average, the more years of education, the higher the level of abstract systematic thinking in a population (all else being equal...).
Abstract systematic thinking then improves our adaptation to the modern world, which improves health and life expectancy in a variety of ways (greater deferred satisfaction, less discounting, more rational behaviour, better salaries from having market-valued skills etc.).
However, abstract systematic thinking does not at present improve reproductive success - but this could also change if/ when educated people become rationally convinced of the benefits of having more children.
Posted by: Bruce G Charlton at Jan 3, 2007 10:28:07 AM
The Lleras-Muney ReSTUD paper (Jan. 2005) that uses compulsory education legislation to examine the effect of schooling on life expectancy looks interesting. I've heard about this paper several times over the last year, but haven't read it yet. Probably need to. It appears that that paper, and subsequent papers looking at other countries' compulsory schooling laws, finds significant and large effects of schooling on life expectancy, and gets us closer to establishing causality and not mere correlations. The Grossman model of health capital predicts such a relationship if I'm not mistaken.
Posted by: Jason Voorhees at Jan 3, 2007 10:41:15 AM
Bruce notes the way that education could effect health outcomes. There's also the fact that improved health involves costly absorbtion of information (see Posner's post on trans-fats). Education will lower those costs, presumably, and thus lower the costs of health production.
Posted by: Jason Voorhees at Jan 3, 2007 10:44:10 AM
Several of the "fat doesn't matter for death rates" studies seem to ignore-- or at least have trouble taking into account-- that sick people frequently start losing lots of weight years before passing away from their illnesses.
I haven't looked up the source studies yet, but I came away from Campos's article with the impression that many of the studies he discusses either categorize subjects by their weight at the start of the study (thus weight loss as a result of a fatal illness will, in the strong majority of cases, not count), and/or only examine pre-elderly mortality, which would tend to overemphasize causes of mortality associated with overweight (e.g. heart attack), and underemphasize chronic diseases where death is likely to be preceded by wasting. Indeed, the latter type of study, while easier from the perspective of experimental design (don't have to track the subjects indefinitely), would be doubly unfair to proponents of the overweight-is-good-for-you theory, because much of that benefit seems likely to show up in old age, both by increasing the chances that illness-induced wasting won't be fatal, and by decreasing the chance that a fall will cause a broken hip. Perhaps I'm mistaken on this point and only the studies he discounts fall into those categories.
In any case, my readiness to believe that the medical CW is completely wrong on this one is not because I have a personal stake in the fat acceptance fight--personally, I'm the always-skinny-and-doesn't-exercise-enough type this argument makes look bad. It's that on the health issues I *do* pay attention to--from the effects of daily moderate alcohol intake, to promotion of a low-fat diet, to the demonization of saturated fats (and promotion of trans-fats as a substitute, haha), to the inane attempts to enforce reduced-salt diets on the public at large--the medical establishment has been not only always wrong, wrong for decades, wrong in the face of clear contravening evidence from their own studies, or wrong by drawing completely unsupported conclusions, or wrong by confusing causation and correlation, or wrong by failing to design studies that actually address the most promising areas for progress, but they are always doing so with a particular bias toward the puritano-masochistic notion that the least pleasurable activity is always and forever the most healthy.
Hence I would not be at all surprised to learn that, as Campos claims, none of the studies purporting to connect high BMI with increased mortality appear to have taken the breathtakingly obvious step of controlling for diet or exercise or cardiovascular fitness. But I suppose I should try to track down some of these studies myself.
Posted by: Dave H at Jan 3, 2007 11:43:00 AM
the studies he discusses either categorize subjects by their weight at the start of the study (thus weight loss as a result of a fatal illness will, in the strong majority of cases, not count)
Depends on if it's a longitudinal study and how long it runs. Some of the studies don't run long enough, and the claim about weight loss as a result of illness not counting doesn't hold. Certain chronic illnesses can last quite a long time.
and/or only examine pre-elderly mortality, which would tend to overemphasize causes of mortality associated with overweight (e.g. heart attack), and underemphasize chronic diseases where death is likely to be preceded by wasting.
Well, don't forget that diabetes is a chronic disease associated with being overweight-- and many theories for type 2 have it being caused by too much abdominal fat. Cancers can also strike pre-elderly as well, and tend to be wasting diseases. So do many of the debilitating diseases that are thankfully rarer now, like TB.
A handful of studies actually do it the way I'd prefer-- a longitudinal study that looks at weight during, say, 30-45 or 30-50 and then compares mortality rates all throughout old age to eventual death of all subjects. Of course, doing a longitudinal study like that requires a long time, so it's rare. However, the one or two studies like that I've seen suggest that, yes, being overweight when 35-45 does increase your mortality rate throughout your life.
I expect that since a proper longitudinal study takes an extreme amount of time and commitment on the part of researchers and patients, most studies will be flawed in one way or another. The meta-analysis is somewhat suspect since some studies are apples and others oranges.
Posted by: John Thacker at Jan 3, 2007 1:21:24 PM
>The main point of the article is that education is strongly correlated with better health outcomes, although the author too quickly assumes a causal connection. For instance education may signal rather than cause low time preference and thus responsible behavior.
This isn't responsive to the paper discussed in the article where the decision to go to school or not was decided by individual states rather than by students themselves.
Posted by: joeo at Jan 3, 2007 4:05:04 PM
Well, don't forget that diabetes is a chronic disease associated with being overweight-- and many theories for type 2 have it being caused by too much abdominal fat.
A criticism I seem to remember of Campos's book was that by focusing so much on mortality rates he ignores the health care costs and lowered quality of life inflicted by Type 2 diabetes. I would be curious to see the extent to which diabetes is actually correlated with BMI (or abdominal fat, or whatever) after controlling for diet, exercise, and genetics, but obviously the diabetes epidemic has gone hand in hand with increasing numbers of overweight and obese Americans and is a tremendous health concern.
a longitudinal study that looks at weight during, say, 30-45 or 30-50 and then compares mortality rates all throughout old age to eventual death of all subjects
Agreed that this is both the ideal way to design the study and logistically onerous. Another concern is that the studies Campos refers to seem to be normalizing based on achieved cardiovascular fitness, rather than, say, duration, frequency and strenuousness of exercise. Maybe the low-mortality, high-BMI types he's talking about really are just those with athletic builds and the actual fat people really are as unhealthy as they seem.
What I'd like to see, then, are studies comparing BMI (or body fat%, or something similar) to mortality rates, normalized for diet and exercise. And I'm particularly intrigued by the suggestion that dieting is counterproductive and actually raises mortality rates. He says the study involved only considers intentional weight loss, so wasting due to disease should not be a factor. Again, I'd like to look up the paper(s) when I have the time.
Posted by: Dave H at Jan 3, 2007 4:57:01 PM
Every 11 year old schoolchild in Scotland was given an IQ test in 1932, and longetivity and health have been found to correlate with IQ, not surprisingly (search for Ian Deary). So, the question is whether education has some effect on health independent of IQ. I would imagine it does, but it's tiresome to see the 800 pound gorilla of social science variables, IQ, ignored for the 1,000,000th time.
Posted by: Steve Sailer at Jan 3, 2007 5:07:19 PM
Dennis Mangan points to Linda Gottfredson's important paper, 'Intelligence: Is it the epidemiologists' elusive "fundamental cause" of social class inequalities in health?'
http://www.udel.edu/educ/gottfredson/reprints/2004fundamentalcause.pdf
Posted by: Steve Sailer at Jan 3, 2007 5:38:02 PM
Steve, you must be great at parties. No matter what someone says, I'm sure you can find a way to bring it back to IQ.
Posted by: Jason Voorhees at Jan 3, 2007 6:11:36 PM
Important as iq is, I don't see how it is relevant to the education studies dealing with sudden changes in compulsory schooling. So yes, the key papers all try to deal with the effects of education on health independent of iq, income, race, etc. If suddenly, all kids born in California in 2005 were forced to take an extra half year of school and that succeeded in lengthening life, then it would seem reasonable to suppose that schooling is a true independent variable -- at least in this very narrow context. And that is akin to what the research tried to demonstrate using early historical data, thus making causation somewhat more plausible though not necessarily conclusive.
I think many of the commentators are missing the point of the econometrics being cited.
Posted by: ghost at Jan 3, 2007 6:19:48 PM
"Steve, you must be great at parties. No matter what someone says, I'm sure you can find a way to bring it back to IQ."
It's called having a market niche. You look around for a useful field that is underserviced by others for irrational reasons. Perhaps economists have heard of the concept?
Economists have strenuously gone out of their way to ignore the vast amount of data on IQ, for reasons largely of careerist cowardice. IQ isn't a panacea for understanding how the world works, but it's the single most underexploited variable in all the social sciences.
Posted by: Steve Sailer at Jan 3, 2007 6:25:42 PM
'Is it worse to be overweight, or not to exercise much?'
I think there may be a non-linearity of sorts here i.e. being excessively overweight is bad and in many cases it is clearly beneficial to exercise, but too much exercise can also be bad.
If the starting point is a person who is at a healthy weight and who is faced with choices between piling on a little more weight or going to the gym, then the choice is more difficult. One problem relating to 'time' is that eating is often enjoyable at the time but regretable later on, while going to the gym can be painful in its anticipation but people usually feel for having exercised both during and after the event. Personally, I force myself to go to the gym even when don't feel like it, knowing that I will feel better for it afterwards. The other issue relating to time is that time is scare and valuable, and going to the gym and showering afterwards takes quite a bit of time, while eating can take a few minutes to an hour or more, depending on your mood.
Posted by: Caravaggio at Jan 4, 2007 8:27:16 AM
THE MOST IMPORTANT THING OF ALL TO
REMEMBER:
Meat is meat. Whether it be fish, pork,
beef, fried lice, chicken, or shark.
Now you remember that!
Posted by: Joe Bleau at Mar 25, 2008 10:57:47 PM
Please come to Anarchy online gold, we will give you a great surprise.
Posted by: Anarchy Online credits at Dec 31, 2008 1:37:24 AM
