Mountains from molehills...

Posted today to the New England Journal of Medicine, is a special report on A Potential Decline in Life Expectancy in the United States in the 21st Century. (Olshsky et. al). The report opens by citing current thoughts in life expectancy increases. Some say we will reach a life expectancy of 100 by 2060, another by 2300.

The Social Security Administration has it's own estimates of course, yet a recent panel has advised the SSA to project longer life expectancies earlier in the future. In case I have to spell this out: Longer life expectancies equate to higher costs, or at current tax rates, less solvency. The Bush administration would love for social security to look as expensive as possible. It is currently politically fashionable to predict people will live longer because it exacerbates our crisis. Here it comes, our worst case scenario....we all live to 100. Fear the day.

So maybe the old actuaries who set up social security were wrong (and maybe they weren't: they've been very good so far). We may have to increase our expenditures on social security.

But there's good news from Olshsky et. al! The life expectancy estimates don't need to be moved up, they should go down. And why down? Because America is getting fatter! Back in the 60's and 70's people were hip and fit, but all throughout the selfish 80's and sluggish 90's obesity was on the rise, with a 50% increase each decade. Now it's so bad that 2/3rds of us are overweight and roughly 1/3 are obese. Look to your left, look to your right: if they aren't obese, then it's probably you. But the good news is you won't be poor from taxes, because everyone will get fat and die young.

There's a few good things to notice about math: the rate of increase in obesity has to drop soon because at the current rate every American adult will be obese by 2035. If you think I'm being farcical, you get a cookie. I disagree with much of the science behind this, but I'll just toss one example out:

Obesity is calculated by body mass index (BMI) which is your weight in kilograms divided by your height meters squares. I gotta say this index trips me out. A freshman in engineering could tell you that isn't dynamic scaling. Consdier 2 people: one of whom is twice as tall, twice as thick, and twice as wide as the other. The large person should have 8x the mass. But the BMI would double even in a proportionately fit person. (I guess the medical argument is people don't scale dynamically). You might also note that people tend to shrink and become more cylindrical as they age...even at a constant weight a little shrinkage puts a hurting on one's BMI.

Anyway, the paper has a few probabilistic models and graphs that assist in it's value. Let's suppose it's good science. The conclusion? Overall life expectancy is 0.3 to 1 year lower than it would be if no one were obese. All this hubub over less than a year? Yeah way to go. The authors point out that this is not negligible, the cumulative loss of life is more than all accidental deaths combined (car accidents, homicides, drownings...etc). But seriously, I don't see a persusive argument that life expectancy is going to drop. I don't even see a persuasive argument not to help myself to seconds.