Picture if you will a time and space where markets were few and far between, big data was unheard of, and national income accounts were not yet a gleam in Simon Kuznets's eye. You will have entered “the statistical dark age.” Or so Rod Serling might have said about the United States before 1840 and especially colonial North America. Into this void have stepped two distinguished economic historians, who have discovered more data than many of us believed were there (although not as much as they would have liked) and have produced “[A] new history of American incomes [that] allows us to take a fresh look at the history of this country's leadership” (p. 250). That history is in two parts: the period from 1700 to 1870 and the period since 1870.

In the latter, they provide a very good synthesis of the literature about economic growth and income inequality since 1870. They delineate three episodes: 1870 and 1914 during which U.S. per capita income once again overtook that of the United Kingdom, while the income distribution remained largely unchanged (Chapter 7); WWI to 1973 (or thereabouts), during which a great reduction in inequality took place, with a decline not only in the top income shares but a leveling throughout the income distribution (Chapter 8); and after the 1970s when rising inequality occurred throughout the distribution, not just at the very top, although it was more pronounced there (Chapter 9). In the most recent period they emphasize the growth of the financial sector, combined with reduced restraints on investment banking, and deregulation of financial markets, as the cause of increased inequality at the high end (p. 239). One of their important conclusions is that U.S. growth was not stimulated by the increasing inequality that occurred since the 1970s.

It is in the earlier period, the statistical dark age, where the authors' original contributions lie. Their idea to estimate the nation's aggregate economic activity from the income side, which provides evidence on both total income and its distribution, is by itself a significant contribution. To carry out their strategy they used social tables to make estimates for five benchmark dates: 1774, 1800, 1850, 1860, and 1870. If those estimates hold up under future scrutiny, they will have some remarkable results.

Their estimates of the income distributions for these years are the first that cover the entire nation and permit comparisons across regions, as well as income classes. There are a number of key findings. First, income inequality in America in 1774 was much lower than in England “or anywhere else on the planet” (p. 95), and the distribution probably had not changed much before the Revolution. Nor was there much change in the distribution of incomes between 1774 and 1800, but from 1800 to 1860 the nation experienced a large rise in income inequality—the Gini coefficient rose from 0.44 to 0.53 for the original 13 colonies—driven by a host of reasons such as increased inequality in earnings, increased inequality in property holdings, and increased urbanization (p. 136). Finally, during the Civil War decade, inequality for the nation changed hardly at all despite an enormous leveling of incomes in the South brought about by the emancipation of the slaves. That was offset nationally because inequality in the North rose even faster than it had in the preceding decades.

Their estimates of national income and rates of growth yield even more striking results, in part perhaps because there are other estimates against which they can be compared. Peter H. Lindert and Jeffrey G. Williamson find that in 1774 average colonial incomes were much higher than those in England and Wales (p. 42). They do not construct a social table for an earlier benchmark date, but argue that since there was little or no growth before the Revolution, in 1700, perhaps earlier, colonial incomes were as high as in 1774, and even higher relative to those England than they were in 1774. Although they do not know how much incomes fell during the war and the period of Confederation, they estimate that the 1800 real income per capita was 20 percent below that of 1774. Since it is widely held that there was some recovery after 1790, the decline between 1774 and 1790 was even larger, perhaps as large as in the Great Depression of the 1930s. On the other hand, from 1800 to 1860 the economy experienced rapid economic growth, racking up an annual average rate of growth of real income per capita of 1.4 percent (the original 13 grew at 1.5 percent per year). The United States became “one of the first, if not the first, to join the modern economic growth club” (p. 103), by which they mean it met Kuznets's criteria of having achieved growth in income per capita in excess of 1 percent per year for a long period of time. The Civil War brought this economic success to a halt. Income per capita for the nation declined only slightly, but the South experienced much larger losses: down 26 percent in the South Atlantic. In general, the South rode this roller coaster history of income per capita to the extreme, due largely to the devastation of both the American Revolution and the Civil War.

Lindert and Williamson have done a marvelous job of explaining how they constructed these new estimates. For each benchmark year they carefully lay out the sources of data, discuss some of their limitations and shortcomings, make clear the key assumptions used in the calculations, provide some sensitivity assessments of the results, direct the reader to the appropriate appendix, and direct the reader to a website with additional electronic files containing the data used. If you try to wend your way through the description, appendices, and links to the data, you will marvel even more at their mastery of all this detail.

As clear and plausible a case as they have made for their estimates, I would caution against a hasty acceptance of them as the new statistical record of American economic growth. Let me summarize their findings in a different way so you can see why I have some skepticism. Lindert and Williamson claim to have overturned just about all the conventional wisdom about U.S. economic growth up through 1870: for example, Robert Gallman's estimate of GDP for 1860 was off by 26 percent (p. 97); Angus Maddisson's dating of the United States overtaking England was off by two centuries (p. 71); John McCusker was wrong about colonial economic growth (Chapter 3); Paul David and others underestimated growth between 1800 and 1860 by 20 percent or more (p. 103); and, of course, income per person in America was already way ahead of that in Britain by the late seventeenth century (p. 74). Lindert and Williamson offer some thoughts on why their estimates differ so much from previous ones, but not enough to be entirely convincing. Here are two examples.

They suggest that their estimate of nominal GDP in 1860 is above Gallman's (which they cite as Weiss's broad GDP estimate) because of a problem with the price deflators. (p. 97) But I do not think this is the source of the difference because Gallman did not use production indexes to extrapolate the 1840 estimates forward to 1860 as they claim. Instead, his estimates were built up using 1860 census figures on the output of manufacturing and agricultural industries, all carefully assessed, adjusted where necessary, and valued in current prices. I suspect that the difference in the two estimates is conceptual. Gallman's GDP figure measures what was actually produced, whereas Lindert and Williamson's is a hypothetical figure of potential output, namely how much could be produced if all resources were fully employed at their most productive uses, and all capital was ascribed its full opportunity cost.

Another issue of concern is the estimate of farm operator profits. The authors point out that their estimate for 1774, taken from Jackson Turner Main's The Social Structure of Revolutionary America (Princeton: Princeton University Press, 1965: 105–14) might include the wage income of the farm owner. They provide an alternative estimate taking that into account, which reduces total income by 12 percent, more so in New England than elsewhere. In the end, they opt to stick with their original interpretation, but I think many people could read Main and come to the opposite conclusion; and might even believe that the profit figure includes some of the return now included in property income. This income component is also a problem for the 1800 estimate because they were unable to make an estimate for that year. So some fraction of the decline in income per capita between 1774 and 1800 is a statistical problem, not a reflection of economic conditions.

In any case, this is an illuminating and provocative book that should garner a lot of attention, and which I expect will inspire much discussion and research. It should also stimulate interest in the colonial economy and in the important, but still relatively neglected, period of transition after the Revolutionary War. The book may not provoke the tsunami of work unleashed by Time on the Cross, but it will likely prompt more research and dissertations than the average book.