Never Enough Numbers
Robert J. Samuelson on the usefulness of statistics
HISTORICAL STATISTICS OF THE UNITED STATES:
Edited by Susan B. Carter, Scott Sigmund Gartner, Michael R. Haines, Alan L. Olmstead, Richard Sutch, and Gavin Wright. Cambridge Univ. Press. 4,489 pp. $825 ($990, beginning in Nov.)
The great appeal of statistics is that they tell stories. Consider these numbers from the latest Historical Statistics of the United States:
From 1960 to 1995, the number of students attending Catholic schools (elementary and secondary) dropped more than 50 percent, from 5.25 million to 2.49 million. In those same years, membership in Catholic congregations increased 43 percent, from 42.1 million to 60.3 million.
Since 1900, U.S. farmers have more than tripled wheat production per acre to 40 bushels in 1997, up from 12. For corn, the gains have been even larger—127 bushels per acre in 1997 versus 28 in 1900. But in the previous century, crop yields barely improved at all. In 1800, wheat yields were 15 bushels per acre and corn yields 25 bushels per acre.
In 1890, the average U.S. tariff on all imports was almost 30 percent. On those imports on which tariffs were actually levied (some goods weren’t subject to any tariffs), it was about 45 percent. These rates typified the 19th century. By 2000, the average tariff was 1.6 percent, and, on dutiable items, 4.8 percent.
From the end of the Civil War to 1900, Americans experienced persistent deflation. From 1865 to 1900, the overall drop in prices was 48 percent, or about 1.4 percent annually. The price of wheat dropped from $2.16 a bushel to about 70 cents.
Now ponder the stories in those numbers.
Catholics have traditionally run the nation’s largest sectarian school system; its decline suggests that, despite an apparent religious revival, the influence of religious schools is waning. The wheat and corn numbers indicate that technology (better seeds and more fertilizers, pesticides, and tractors) explains the 20th century’s explosion of food production; previously, the expansion of farmland was the main cause. High tariffs in the 19th century contradict the notion that free trade aided America’s early economic growth—though it may aid economic growth now. Economists sometimes express fears about deflation, but modest deflation historically has not been crippling. The economy was four times larger in 1900 than in 1865.
This is the first edition of Historical Statistics since 1975. The Census Bureau, which had published the three earlier editions beginning in 1949, didn’t receive sufficient congressional funding to continue doing so. In 1995, a group of scholars headed by economic historians (and wife and husband) Susan Carter and Richard Sutch of the University of California, Riverside, took up the job. In the end, 83 scholars contributed their tables and time in return for a copy of the finished product.
The new edition is a monster—five volumes, versus two for the 1975 edition. But that’s understandable. By some estimates, more than four-fifths of the scholarly historical data series have appeared since 1970. New topics include poverty, American Indians, and the Confederacy. Many of the statistics are eye opening. For example, from 1945 to 1995, the number of guns per capita nearly tripled, from 35 per 100 people to 92. But the share of homes with a gun decreased, from 49 percent in 1959 (the earliest year for which data are provided) to 40 percent in 1996. Apparently, guns are like TVs: People who have them have more of them.
Omissions? Well, yes. Public-opinion polling data are almost entirely absent. And there’s nothing on sex (though statistics do exist). But the set’s biggest defect is its price—$825, rising to $990 in November. Statistics fanatics will probably be able to find copies at many libraries. Universities and colleges likely will buy the online version for their faculties and students. Still, here are a couple of better ideas for the publisher: How about a one-volume abridged edition at $75? Or a CD-ROM of the full set for $250?
—Robert J. Samuelson