bell curve?

Date view Thread view Subject view Author view

From: Dave Long (dl@silcom.com)
Date: Thu Oct 12 2000 - 13:53:46 PDT


> Sure, as long as the big fatty part of the bell curve gets big and moves
> to da retro. "Wide Right," call it Scott Norwood economics. The area
> under the curve isn't a constant over time. I would much prefer a big
> distribution centered around $74,000 with the sigmas getting shafted than
> a real nice, tight spike where everybody sits around $12,000. Too bad you
> can't create any artificial plateaus.

What bell curve? Neither wealth nor income are normally distributed,
at least according to the figures available at the census bureau.
I suspect lambdas may be better than sigmas for talking about them,
but maybe taking log($) would suffice.

Household net worth (ca. 1993) had a mode at $1-$4999, and fell off
with a long tail to the right. (12% < 0, 0 < %14 < 5k, 500k < 4%)

Personal income (ca. 1999) for the 70% of those over 25 who had
earnings actually rises a bit from $0-$2499 (5%) to a mode at
$20000-$22499 (6%), and falls off from there. It's about 1% per
$2500 increment around $70k, and a quarter of a percent per over
the area from 85-100.

So, from the standpoint of income, $22k is better than $12k, but if
we were talking about wealth, $5k is worse. Either way, the skew
means there isn't much centering going on.

-Dave

1999 Income, M+F, 25+, $k

none ####################

  10 ########
  20 ##########
  30 ##########
  40 ########
  50 #####
  60 ###
  70 ###
  80 |
  90 |
 100 |

  oo ##

1993 Net Worth, Household, $k

< 0 ###

  20 #######
  40 ###
  60 ##
  80 ##
 100 #
 120 |

  oo ######


Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Thu Oct 12 2000 - 13:51:30 PDT