Published by EH.NET (October 2006)


Susan B. Carter, Scott Sigmund Gartner, Michael R. Haines, Alan L. Olmstead, Richard Sutch, and Gavin Wright, editors, Historical Statistics of the United States: Millennial Edition (Online Version). New York: Cambridge University Press, 2006.

(Note: The price of Historical Statistics is going up on November 1 — from $825 to $990 for the five-volume print version, and from $1,250 to $1,450 for the online version purchased by individuals. The online version for libraries is more expensive and the price depends on the library’s size.)

Reviewed for EH.NET by Samuel H. Williamson, Department of Economics, Miami University.

The publishing of Historical Statistic of the United States: Millennial Edition is a major event for American economic historians. It comes in five volumes and is also accessible online. The online package will be reviewed below, but first I want to comment on the entire project.

EH.Net has already published a review of each volume and all the reviews have been full of praise. Here is a list of the volume titles, each of which is linked to its review.

1. Population —

2. Work and Welfare —

3. Economic Structure and Performance —

4. Economic Sectors —

5. Governance and International Relations —

The work, however, is more than the sum of its parts. This is the fourth version of Historical Statistics (HS) and it is greatly expanded from the previous editions. It is also the first edition not published by the U.S. Census Bureau.

The first HS came out in 1949 as an appendix to the Statistical Abstract of the United States. The second edition was published in 1960, followed by the Bicentennial Edition in 1975. The first three editions had 3,000, 8,000 and 12,500 series, respectively. The current edition has 37,339 series that include dozens of new topics and introductory essays explaining the sources and limitations of the data presented. One of the appendices (found in volume five), written by Carmel Chiswick, explains the origins of HS and is very useful for putting the current edition in perspective.

This work has been twelve years in the making. It started in 1993 when a group of interested organizations met with representatives from the Census Bureau who indicated that they could not commit to a new edition, but would support the venture if these groups did it. In the end, over 80 scholars with the support 70 plus organizations produced this Millennial Edition. While it was a huge project that depended on the contributions of many, in my opinion, two people should be especially singled out. They are Richard Sutch and Frank Smith. Sutch organized the original meeting with the Census Bureau and (with no disrespect meant for the considerable work of the other four editors) based on my observations, he has been the chief organizer ever since. If I understand correctly, Smith has been the economics editor from Cambridge Press in charge of this project over the past twelve years — the one who has had to bear up under missed deadlines and unanticipated expenses and yet who ultimately saw the project through to the valuable, polished finished product we are enjoying today.

The delays in the project will be readily evident to most readers. For example, a large number of series are from census data, but most of the edition’s data collection was done before the 2000 census data were available. Thus, most (but not all) of the decadal series end in 1990. Many of the annual series end in 1998 or 1999, though a few extend as far as 2002. There are, however, many, many series that end at other years. For example, the twenty-seven tables on employment and industry end in eight different years from 1940 to 2000.

This quibble aside, it is important to note that each table comes with source notes, documentation and definitions that are very helpful. However, while it is understandable the editors wanted to include so many useful series, this inclusiveness may overwhelm the average user.

My review of the online version has been delayed partially because this version has been a work in progress. When I first started, there were too many programming glitches for me to give the work a fair review. However, the editors at Cambridge Press have been very open in communicating with me and in working to fix these problems. Perhaps the biggest advantage of an online version would be the ability to copy tables and paste them into a spreadsheet. However, when I initially tried this, I found that negative numbers were not formatted to be able to paste. This problem has recently been fixed.

The online version has all the tables and essays that are in the print version and reports all data in an unrounded format. There is a good search engine that is helpful in finding things, which is particularly useful if the user does not have the print version available. In addition, the version allows users to create custom tables that can combine series from all the diverse series.

The most important tool is the graphing program. By creating a custom table, the user can compare any of the series over time or in a scatter diagram. For a user accessing this program online in a library, this will be an invaluable part of the package. The user can create graphs in color and print or download them. The online user can also send tables to any email address and the recipient has ten days to view it. While the program notes that “you can download [output] in PDF and/or Excel during this time,” this appears to be only true if the recipient has an account. On the other hand, the numbers in the tables can be copied from the browser.

My tests of the graphing tool were 1) creating a custom table and then using the graphing program; 2) creating a custom table, downloading it and then creating a graph in Excel; and 3) going to the various series, copying the numbers I wanted, pasting them into Excel and creating a graph. For me the third choice won, but I must add that I am an experienced Excel user and I could not have done this if I were not logged on to HS on a computer that also had Excel. Most researchers are going to want to use Excel or some other package to graph, as the program on HS has fewer graphing options.

The Government Printing Office still sells the 1975, Bicentennial Historical Statistics edition, which is 1,200 pages in two volumes, for $114.50. The Millennium edition of five volumes, with 4,500 pages is selling for $990 as of November 1, making it more than eight times as expensive. This price may mean that these volumes will seldom be found outside university and other research libraries. With the online version at $1,450, it will even be less likely to be commonly found. On the other hand, everything but the tables and essays are available free online, so that a user can search for a table or essay and then view it for printing or downloading for $6 each

I have no idea what percent of the 37,339 series are currently still being collected and could be updated. For those that have an unchanged definition and are from public sources, such as the Census, it should be quite easy. For others that are period specific or where a base year has been updated, or the bundle has changed — such as real GDP or the various price indexes — it would be difficult or just not practical since often these revisions are to all the observations.

The online version of Historical Statistics? is a great innovation, but it would be much more valuable if it was more than a reprint of the text version and contemporary tables were updated. I would expect that one full-time researcher could update dozens of them a day. Integrating them into the format of the tables should be possible.

My other suggestion to the editors is that they publish an HS “light” with a few hundred of the most popular tables and link it to the online version. If it could be priced for less than $100, then small libraries and others could buy it and would be directed to the online version to purchase the essays and other tables for the $6.

Samuel H. Williamson, cofounder and president of MeasuringWorth (, is Professor of Economics, Emeritus, from Miami University. In 1983 he was a cofounder of the Cliometric Society and served as its executive director for sixteen years. In 1996 he founded EH.NET and was its executive director until 2003.

Copyright (c) 2006 by EH.Net. All rights reserved. This work may be copied for non-profit educational uses if proper credit is given to the author and the list. For other permission, please contact the EH.Net Administrator (; Telephone: 513-529-2229). Published by EH.Net (October 2006). All EH.Net reviews are archived at