Iam grateful to Alice Amsden, Martin Bronfenbrenner, Christopher Clague, David Felix, Joseph Reid, Kazuo Sato, and referees of this Journal for insightful comments on an earlier draft of this paper. I also thank the Faculty Research Program of the Columbia Business Schoolfor financial support, andjohn Millarfor able research assistance. I bear sole responsibility for any deficiencies in the paper.

Journal Article The Stagnation of Brazil's Exports: Reply Get access Nathaniel H. Leff Nathaniel H. Leff Columbia University Search for other works by this author on: Oxford Academic Google Scholar The Quarterly Journal of Economics, Volume 86, Issue 3, August 1972, Pages 508–510, https://doi.org/10.2307/1880810 Published: 01 August 1972

I. Data on export stagnation, 286. — II. Causes of export stagnation. 287. — III. Effects of export stagnation, 292. — IV. Possibilities for change, 296. — V. Conclusions, 299.

I. Emergence of the disparity, 244. — II. Differences in rates of regional export growth, 245. — III. Some questions, 248. — IV. Earlier explanations, 249. — V. An interpretation of the decline of Brazilian sugar and cotton exports, 252. — VI. Land and imperfect factor reallocation, 255. — VII. The exchange-rate mechanism under imperfect factor reallocation, 255. — VIII. Currency-area effects and regional divergence, 258. — IX. Conclusions, 260.

Arthur Goldberger's comment is a tempest in a teacup. This is readily verified from the data presented in Table 1. This table reproduces from my original paper the coefficients for paired equations estimated with three different samples in which In S/N and ln SI Y were the dependent variables. As the data of Table 1 indicate, in 7 of the 12 cases, the coefficients are identical through the second decimal place. Furthermore, all the discrepancies are quantitatively very small. In 9 cases, the paired parameter estimates are within 1 percent of each other. In the remaining cases, the differences are of the magnitude of .0825, .0012, and .0011. These are well within the margins of accuracy within which anyone views data generated from the national income accounts of the less developed countries. In light of these considerations, it is difficult to understand the tone of Goldberger's comment. The reason for the slight discrepancies is straightforward. My observations for S/N and S/Y were computed separately from data supplied by the Statistical Office of the Agency for International Development. With the rounding introduced by the series for N and Y, it is not surprising that the series for S/N is not identically equal to the series which would be obtained by forming S/N from the product of Y/N and S/Y. Similar slight discrepancies have also been reported in other econometric work in which equations for S/N and S/Y have been estimated with separately computed data series.' In any case, nowhere in my paper did I even refer to the precise numerical parameter estimates obtained in my equations. The main conclusions of the paper were that the dependency variables were quantitatively an important determinant of international savings rates, and, further, that introduction of these variables greatly reduced the importance of percapita income, on which some previous discussions had focused. Neither of these conclusions is affected by Goldberger's comment.