Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

The values you're quoting are wealth per adult, which is a mean, not a median, and therefore very useful to Credit Suisse in context and almost entirely useless in the current discussion.

To get a better feeling for how things look in terms of the actual humans involved, a median isn't bad, or a percentile based distribution, or you can e.g. cut off the top and bottom 10 or 20% and then do some sort of analysis based on that.

'Mean net worth' is only useful for a very specific set of situations - it's also why you often get charities talking about 'look how tiny the net worth here is' but that includes people with debt having negative net worth, which drags the numbers downwards in just as distortionary a way as the richest people in the US are dragging the numbers you're using upwards.

Of course, it makes great propaganda for the charities, and they're not -strictly- lying - and Credit Suisse aren't lying at all, they're presenting total wealth by country and providing the 'per person' number to help gauge that information, not for use as remotely representitive.

However if you keep scrolling down the document, you'll find Table 2 provides the median - and lo and behold because the mean for the US is skewed by the small (compared to total population) number of extremely rich people, the US in No. 18 by median wealth as opposed to No. 2 by mean.

Statistics are great fun but you do have to be careful to apply the right -sort- of statistics for the purpose you have in mind.



I gave mean and median




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: