Very few issues in American economic history have stirred as many contrasting views as the debate regarding the effects of the British Navigation Acts on the Thirteen American Colonies. The most recent entry in the long argument is Robert Thomas' analysis of the problem. Thomas constructs quantitative estimates of the impact of the Acts which are designed “… to measure, relative to some hypothetical alternative, the extent of the burdens and benefits stemming from imperial regulation of the foreign commerce of the 13 colonies.” He presents these estimates to test the hypothesis that: “… membership in the British Empire, after 1763, did not impose a significant hardship upon the American Colonies.” Thomas has employed new data in constructing his estimates, and he has carefully brought economic theory explicitly into the discussion. His framework requires the construction of a “counterfactual” hypothesis against which to evaluate the impact of policy. As he views it: “The only reasonable alternative in this case is to calculate the burdens or benefits of British regulation relative to how the colonists would have fared outside the British Empire, but still within a mercantilist world.” Given such a model, the value of the burden and benefit is estimated in terms of market prices and the pattern of trade which would have emerged had the colonies been free in 1763.

Professor Sandberg in a very helpful comment has accepted the main conclusions of my paper on die economic rationality of New England mill owners with respect to the adoption of the Draper loom but has raised certain questions as to my handling of some of the data used to reach this conclusion and of the very representativeness of the data themselves.

In a recent issue of this Journal, Ian Drummond took issue with my article discussing the relation of labor scarcity to British observations of American industrial efficiency in the 1850's. I tried to show that (1) the contemporary observations were not specific enough to show whether or not Americans were using the same technology as the British; (2) even if the observations showed that the Americans used the same technology as the British, it would not follow that the higher ratio of land to labor in the United States had induced a higher ratio of capital to labor; and (3) if the observations proved that the Americans used a more efficient technology than the British, the connection between this fact and the high land-labor ratio was not clear. Under (2), I tried to demonstrate that the land-labor ratio and the capital-labor ratio were separate, and that it was improper to make inferences about the latter from data about the former. In fact, if we assume that land was not used in manufacturing, the data imply that while the land-labor ratio in America was higher than in Britain, the capital-labor ratio was lower.

Approximately two years ago Stanley Lebergott took issue with the finding “that the rise of the railroads was relatively unimportant” to the American economy of the nineteenth century (p. 438). Lebergott dis-missed the low estimates of social savings as “startling but uninteresting,” and offered an alternative estimate of railroad importance. He asked the important question “by how much the railroad reduced real costs of transport from those that would have prevailed under a regime of wagoners and boatmen” (p. 441). He approached the answer by way of two other questions. Given cost and technological data, “what was the maximum volume of goods that could be distributed per year by a mile of canal and by a mile of railroad where both could reasonably be built? And how great was that volume per dollar of investment?” (p. 422). Using nineteenth-century data in his inimitably artistic manner, Lebergott fashioned an answer suggesting that the consequences of railroad construction were “surely the stuff of significant economic change” (p. 466).

A factor of proportionality was inadvertently omitted from equations (6), (7), (9), and (13) of “The Specification Problem in Economic History,” JOURNAL OF ECONOMIC HISTORY, XXVII (Sept. 1967), 283-308. A tilda was omitted above PA in equation (9).

Roger Ransom takes exception on three counts to my published study of the burden of the Navigation Acts, questioning (1) the counterfactual proposition I employed, (2) my estimate of the relative burden, and (3) the likelihood of differing regional impacts of the Acts. I shall briefly take up each of these in turn.

Is a railroad like a toy wind-up set, the locomotive moving endlessly, stupefyingly around a single circle of track? Or, as I believe, is it more like a fancy set, with various sidings, and connecting routes? Professor Weiss assumes that the carrying capacity of a railroad is bounded by its ability to store all its freight cars on the main line. And the view he (somewhat incorrectly) attributes to me—that “for every starting point there are 10 miles of clear road ahead”—is nothing more than the belief that a real railroad is not a simple circle. For railroad sidings do exist for inventorying cars. And connecting railroads that move cars out of a particular system do exist. Economists who contrast the productive capacity of Ford and GM rarely worry about whether a week's production can be stored within the factory. Entrepreneurs envisaged slack variables before they were baptised as such. (If sidings did not exist, or were a significant portion of the cost of one transport system but not another, matters might be quite different.)