Knowledge that Transforms

I am grateful to Professors Ralph Andreano, Arthur S. Goldberger, and W. Lee Hansen of the University of Wisconsin for their encouragement and comments on an earlier draft of this paper. Many valuable suggestions were contributed by two anonymous referees and the participants in the Workshop in Applied Welfare Economics at the University of Wisconsin, Fall 1970, especially Morley Gunderson. Charles Leitzke assisted in the arduous task of data collection. The usual caveat is applicable.

AS CRITICISM of the traditional theory of production and exchange has mounted in the postwar period, increasing attention has been given to new analytical approaches that seek either to supplant classical marginalism or to extend its scope. In the latter category is the important body of literature that has grown up around the notion of property rights structures. The contributions here are quite diverse in style and content but are characterized by a common emphasis on certain basic ideas concerning the interconnectedness of ownership rights, incentives, and economic behavior. The purpose of the present paper is to summarize the essential features of this line of research, examine some of its important areas of application, and discuss the promise the approach holds for improved understanding of economic problems.

Although lack of price data on beer for the past decade makes prediction impossible, we note that from 1962 to 1969 per adult beer consumption increasec1 by 16 per cent (USBA, 1970, p. 59) while real, per capita income rose by 27 per cent. Assuming stable real prices for beer, the implied income elasticity is 0.59. Average per capita income (1958 dollars) during 1962-1969 was approximately $2,250. At this level we estimate income elasticity as 0.37. Since average state taxes on beer increased about 7 per cent in real terms during 1962-1969 (USBA, 1970, p. 105), our estimate appears sufficiently close to warrant confidence in its reliability. Regarding implications for the future, it appears that the beer industry can anticipate moderate increases in consumption, assuming the price of beer remains relatively stable. Moreover, it further appears that excise taxes on beer could be reduced considerably with only modest losses in tax revenues. The desirability of such reductions might be questioned on noneconomic grounds; however, there appears little doubt that beer excise taxes are regressive. REFERENCES

Using the same sample of 100 countries for 1966 as Sommers and Suits,3 we obtained the following equation 4 GCF/GNP 26.87 (1.04) 4676.54 / (GNP/N + 300). (629.25) R2= .4 (4) Comparing this result with the estimation of the quadratic equation (1) of Sommers and Suits,5 we see that both equations have similar statistical properties. Using (4) as the basis for simulations of the growth path, however, we obtain completely different results from those of Sommers and Suits. Now per capita income does not attain a stationary level, but grows exponentially. The growth rate of per capita income, however, as can be seen from figure 2, attains a stationary level at 4.41 per cent per year. FIGURE 2. SIMULATED GROWTH RATE OF GNP PER CAPITA

GCF gross capital formation GNP gross national product N population r growth rate of GNP/N. After fitting this model to a cross section of 100 countries for 1966, they use the estimated coefficients to simulate a growth path of a typical economy. In this comment it will be shown that the simulation results depend critically on the quadratic sDecification of equation (1) .2 Although the estimation of equation (1) by Sommers and Suits gives satisfactory results, there is little empirical evidence for the declining range of the equation. This can easily be seen from the scatter diagram and the graph of the fitted equation (figure 1): GCF/GNP attains its maximum when GNP/N is $2,169. The sample, however, contains only 8 countries (out of 100) with a per capita income of more than $2,169. Except for the single case of the United States (GNP/N $3,763 and

Haitovsky, Y., Multicollinearity in Regression Analysis: Comment, this REVIEW, LI (Nov. 1969), 486-489. Hocking, R. R., and R. N. Leslie, Selection of the Best Subset in Regression Analysis, Technometrics, Vol. 9, No. 4 (Nov. 1967), 531-540. Kendall, M. G., A Course in Multivariate Analysis (London: Charles Griffin and Company Ltd., 1957). Kendall, M. G., and A. Stuart, The Advanced Theory of Statistics, 3 (London: Charles Griffin and Company Ltd., 1966). Lancaster, K., Mathematical Economics (New York: Macmillan Company, 1969). McCallum, B. T., Artificial Orthogonalization in Regression Analysis, this REVIEW, LII (Feb. 1970), 110113. Mitchell, B. M., Estimation of Large Econometric Models by Principal Component and Instrumental Variable Methods, Technical Report no. 28 (Economic Series), (Stanford, Institute for Mathematical Studies in the Social Sciences, Stanford University, 1970). Stewart, J., Letter to the Editor, The American Statistician, 25 (Apr. 1971), 40. Theil, H., Economic Forecasts and Policy, 2nd ed. (Amsterdam: North-Holland Publishing Company, 1961).

IN 1963 Nissan Liviatan tested the Permanent-Income Hypothesis (PIH) as an errors in variables model and used a combined crosssection time-series body of data consisting of the 1958-1959 Israel Reinterview Savings Survey. His findings were weakened (i) by the fact that a savings survey was used which meant that consumption was not independently estimated but rather was computed as a residual; hence, consumption and income had a common measurement error which thus violated a basic assumption of the model that the covariance between transitory income and transitory consumption is zero; and (ii) by the fact that, if the horizon is in fact three years, then income lagged one year cannot be used as an instrumental variable. Milton Friedman, in his rejoinder accompanying Liviatan's paper (1963), suggested that it would be most desirable that a Livia.tantype analysis be applied to data not marred by common errors of measurement and that such data span a three-year period so that income or consumption for one year could be used as an instrumental variable for a year at least two years later or earlier. In this paper both criteria will be met income and money demand are independently estimated and cross-section data covering a three-year period are used. In addition, this paper extends the PIH of aggregate consumption to individual consumer goods. The generalization results from relaxing Friedman's assumption of a linear homogeneous consumption function (unitary income elasticity). Since Friedman has elsewhere (1959) suggested that money can be looked on as a consumer durable, it will be money that this study uses as its consumer good. The methodology, however, is completely general and any other consumer good could have been chosen.