Wednesday, April 16, 2014

On historical statistics

Thanks to this Swedish website on historical statistics, I tracked down the Cambridge University Press published Historical Statistics of the United States (Millennial Edition Online), where I am now a member. The HSUS-MEO have good stuff, but they want $6 a table for a 48h window of usage, so I have to plan my uses carefully.

As an experiment, I looked into the monthly wages for farm laborers that received board (i.e. Table Ba4234-4234 Monthly earnings with board of farm laborers, by region: 1818-1948), and found the following information for the areas I am most interested in.

  • United States (Average?)
    • 1818: $9.45
    • 1826: $8.83
    • 1830: $8.85
  • New England
    • 1818: $11.90
    • 1826: $11.65
    • 1830: $11.60
  • East North Central
    • 1818: $8.86
    • 1826: $8.73
    • 1830: $8.73
  • West North Central
    • 1818: $10.15
    • 1826: $10.15
    • 1830: $10.15
(itself taken from Stanley Lebergott, Manpower in Economic Growth: The American Record since 1800, (McGraw-Hill) 1964, Tables A-23 and A-24, pp.257ff.)

Robert A. Margo, who contributed this information to the HSUS-MEO, writes that
For 1818, 1826 and 1830, estimates were made in 1832 by Senator John Holmes of Maine and reported by him in the Congressional Register of Debates. For certain states there are, in addition, the results of a survey in 1832-1834 on 1832 farm wages made by Secretary of State Edward Livingston, drawing on returns from many individual towns in these states (that is, 59 of 134 towns in Connecticut, 101 of 444 in Main, 109 of 230 in new Hampshire, and so forth). 
Margo used the Livingston series to compute the 1830 value, then used the Holmes series to extrapolate 1818 and 1826 values, using the census information of the 1820 census to compute weights.

The Holmes series should be either in the 1st Session or the 2nd Session of the Congressional Register of Debates.

No comments:

Post a Comment