Did the gap between rich and poor widen or narrow (1) between 1870 and
1970 and (2) between 1970 and 1995? In other words I am looking for
one comparison between 1870 and 1970 and a second comparison between
1970 and 1995. I would like some actual statistics. Could you also
quote the source so that I can verify it and also quote it in an
article I am writing.

Bill --

Thank you for the interesting -- and relevant question!

The U.S. Census Bureau has maintained detail records from 1967
forward.  A top-level view is the "Historical Income Tables --
Housholds," (Sept. 30, 2002) provides distribution by quintiles:
http://www.census.gov/hhes/income/histinc/h01.html

There are a variety of detail tables, including breakdowns by race,
region, state, age and education in this collection of tables:
http://www.census.gov/hhes/income/histinc/inchhdet.html

Earlier Census measures used a CPI-adjusted level of household income.
 The "Statistical Abstract of the United States: has percentage
distributions for seven divisions between "under $10,000" and "$75,000
and over," (pages 21-22 of this Acrobat document):
http://www.census.gov/prod/2002pubs/01statab/income.pdf  

You may be interested in seeing this report "Shares of Income by
Quintiles," done by Census staff member John McNeil (undated) to
discuss the changes in data format:
http://www.census.gov/hhes/www/income/newydata.pdf

Finding income distribution numbers for the late 1800s involves using
other measures.  Even for measures of Gross National Product
"estimates start in 1500, and often are no more than educated guesses,
especially before 1870"
according to the authors of a National Bureau of Economic Research
(NBER) paper, "The Rise of Europe: Atlantic Trade, Institutional
Change and Economic Growth," (March 28, 2002).

Instead, researchers rely on rough GNP numbers and estimates of the
worth of well-known Americans.  A paper by Brad DeLong (UC Berkeley)
relies on several books to estimate "Wealth Concentration in the U.S.:
Share Held by Top 1% of Households."  A chart in the middle of the web
page shows the variance in concentration over time:
http://www.j-bradford-delong.net/Econ_Articles/back_on_top.html

Other DeLong pages attempt to estimate the value of the most-wealthy
Americans.

DeLong's data relies on three books:
"Three Centuries of American Inequality," Jeffrey Williamson and Peter
Lindert,
"Wealth of a Nation To Be," (for the Revolutionary War period), Alice
Hanson Jones
"Men and Wealth, 1850-1870," Len Soltow

NBER papers provide interesting peeks into old historical income
numbers.  While not a complete sample, this report lists income
distribution among Civil War participants:
"Prior Exposure to Disease and Later Health, Mortality: Evidence from
Civil War Records"
http://www.nber.org/books/healthandlabor/lee8-23-01.pdf

Both professors Williamson and Lindert have stayed active in the
discussion of income distributions.  You may wish to contact them
directly for insight into historical numbers.  Lindert is at UC Davis
and Williamson at University of Copenhagen and Harvard.

Google search strategy:
www.census.gov
"U.S. income distribution" + 1870
www.nber.org  + "income distribution"

Best regards,

Omnivorous-GA

---

Request for Answer Clarification by bill666-ga on 20 Nov 2002 05:12 PST

Thank you for prompt and helpful answer -- well worth $25! A niggle is
that you did not attempt a direct answer to my two questions. However,
I suspect that I can assume you think the gap narrowed in the first
period and widened in the second from reading the documents you
referred to, in particular Bradford de Long's paper: "The US Economy
"Back on Top"?". Are you willing to take the plunge and be categoric
on these points?

---

Clarification of Answer by omnivorous-ga on 20 Nov 2002 07:23 PST

Bill --

For an opinion, you'll have to go to Google Opinions.  Believe it or
not, I wanted to see what data was available, knowing that information
in the early periods would rely on some proxies to measure income. 
Early Census records made some attempt at estimating wealth (1850 and
1860) and surprisingly it didn't show in any of the research.  Knowing
human nature, it's easy to suspect that attempts to measure wealth in
the mid-1800s were unpopular and subject to large errors.

What did your article say?

Best regards,

Omnivorous-GA