Hacker News Reader

Comment by jamix | Hacker News Reader

> Life expectancy in Hong Kong is 84.23 years, more than five years longer than the US and the highest in the world. Hong Kong is not that wealthy (median household income is $38,000 USD); it’s somewhat polluted; people don’t obviously eat what seems like a healthy diet; and they don’t seem to exercise a great deal. What should we learn from this?

We should learn that Hong Kong is probably a smaller territory so any of its "average" metrics has a higher chance of being an outlier, out of sheer randomness.

View on HN

One of the comments on the page points out this:

> Hong Kong life expectancy: Hong Kong is a city with no rural area. If you look at the life expectancy of just Sydney, leaving out the rest of Australia you get 86.7 years instead of 82.5. It's no secret that life expectancy in cities is generally longer, especially if they are well to do ones that don't make their money from accident prone industries.

The use of political borders to make regional comparisons is a common source of ridiculous conclusions. It usually happens in more banal and easily-identified ways, like when you hear somebody say something like, "Austin is bigger than Boston," (no, it isn't, obviously) but this shows how it can fool even very smart people if they're not paying attention.

mmm, isn't Austin bigger than Boston? I'm not sure what you mean there.

Austin isn’t anywhere close to the size of Boston, neither statistically nor experientially. That this isn't obvious to everyone sort of proves my point. The use of political borders as boundaries for making regional comparisons only misleads.

I really don’t think I can put too fine a point on this: the Boston region is so much bigger than the Austin region that nobody familiar with both places could possibly think Austin is bigger. The only way to think Austin is bigger is to be mislead by the arbitrarily drawn political borders of each region’s core city. (I’m also taller than Shaq if for some silly historical reason we only define Shaq to include the parts of his body up to his knees.)

It's over 7 million people. That seems like a robust dataset to me.

The fact that it’s a smaller territory means that effects that would even out in larger territories don’t apply. I.e. it’s a relative effect.

Here’s an example: suppose a large retirement community of Chinese immigrants exists in Hong Kong, raising the average age. In order for this kind of retirement community to shift the age distribution the same amount in a country 10x larger, that retirement community would have to be 10x or so larger as well. A 10x larger retirement community is way less likely.

This isn’t to say that all small things lie on the extremes of the distribution, only that an extreme value is likelier to come from a smaller sample.

> That seems like a robust dataset to me.

Perhaps, but there's a wide distribution of age of death, and the author is fretting about a 5 year difference in life expectancy between 2 populations that are doing fine.

Assuming someone can ever fully explain the cause of this life-expectancy difference, it's going to be multi-faceted and complex and amount to nothing more than a long list of generic advice and policy prescriptions that will go nowhere and probably be culturally incompatible.

that's not a winning attitude

Size isn't the salient factor. Consider two "cities," one of which draws its political borders around its inner suburbs and one that doesn't. They may have identical populations, even, but they'll produce very different results on many common statistical measures.

This is why you should almost never compare two places based on political borders or designations. (You have to instead normalize around something less arbitrary. "City" and "state" aren't proper categories for classification, because what gets included from city to city and state to state can vary across multiple dimensions.)

This list seems like a plausible set of causes: https://edition.cnn.com/2018/03/02/health/hong-kong-world-lo...

Given that this is Marginal Revolution, I wouldn't expect them to accept any of the answers that involve state intervention.

There's another possibility: several of the "oldest person in Japan" were found to have been dead for some time and their relatives were claiming their pensions.

Yes just like the "Blue Zone" diet areas were found to be curiously poor parts of countries. It turns out they weren't the places with the most residents over 100, they were the places with the most... poor record keeping and fraud.

https://www.biorxiv.org/content/10.1101/704080v1

From the comments on the post:

Brian Kearns

“ According to a Stanford University study recently published in the science journal Nature, Hong Kong people walk an average of 6,880 steps a day, making us the most ambulatory populace out of the 46 territories and countries assessed.”

Japan has a life expectancy of 83.98 years, and a population > 120 million

They walk a lot too. If you want to live a long time walk as much as possible.

Also in the US there is a race based discrepency for life expectancy. This is what I got for 2014:

* Native Americans: 75.06 years

* African Americans: 75.54 years

* White Americans: 79.12 years

* Hispanic Americans: 82.89 years

* Asian Americans: 86.67 years

I think people in HK get more exercise in general by... walking. Walking here is the key.

My theory would be that life expectancy measures more length of life than quality of life. There's also Asian culture in place.

For example, an old person eats unhealthy, smokes, makes lots of irresponsible decisions in life. She gets cancer, stroke, diabetes, kidney failure, we can still keep her alive for decades. She might be hooked to a dialysis machine. She has little savings. She might not have the money, time, or energy to deal with this.

Asian culture dictates that her only daughter quit her job to take care of the sick mother. There's a huge drain on society, but it keeps someone alive longer.

Reply on news.ycombinator.com