Knowledge that Transforms

This paper examines the cross-industry determinants of the importance of Japanese direct investment activities in U.S. manufacturing through an extension of the analysis in a recent article by Bruce Kogut and Sea Jin Chang (1991). The results indicate significant positive roles for Japanese technology and marketing assets. While U.S. technology assets are insignificant, a significant negative effect for U.S. marketing assets suggests that entry barriers related to marketing are more important than any use of foreign direct investment by Japanese firms to access these assets. In addition, U.S. government policies toward defense-oriented industries appear to act as a deterrent to Japanese direct investment. Copyright 1996 by MIT Press.

This paper uses a survey of risk beliefs about skin cancer to provide new evidence on how people view risky situations. Empirical results presented are based on a measure of risk beliefs held at the time of the survey. Key findings are that risk beliefs about skin cancer account for factors including skin type, complexion, and sunlight exposure history. Also, the connection between risk beliefs and willingness to pay is explored by using reservation prices for a sun protection product. A new method for treating joint production in a household production framework is developed to support this analysis. Copyright 1996 by MIT Press.

The authors test a standard expected utility model and alternative models about how people evaluate risky prospects using data about individuals' preferences among health insurance plans. A model that assumes people evaluate gains and losses relative to a reference rather than final outcomes, treat gains and losses asymmetrically, and process certain and uncertain outcomes separately provides a better fit than the standard utility model. These findings suggest inertia in health insurance plan choice and that individuals are more responsive to decreases than to increases in the price of insurance. Copyright 1996 by MIT Press.

Let x be total consumption expenditures. Measurement errors in consumer budget allocation models cause unusual problems because errors in the dependent variables, consumption on individual goods, contribute to errors in the regressor x. Empirically adequate budget share models, like the AIDS, translog, and quadratic AIDS, are linear or quadratic in log x. It is shown that, for these specifications, two stage least squares fails to correct for the measurement error. A simple technique for recovering consistent estimates from two stage least squares is provided, along with an efficient generalized method of moments estimator. The method is applied to U.K. fuel demand data. Copyright 1996 by MIT Press.

This paper develops and estimates a cost model for U.S. hospitals, analyzing the cost of excess bed capacity. A new estimate is worth making for at least two reasons. Recent changes in the economic environment of hospitals have caused their utilization rates to fall sharply, making previous estimates inaccurate. Second, we employ econometric techniques not previously applied to this problem, with estimation based on all short-term community hospitals from 1979-89. Our results, based on conservative estimates of the average optimal occupancy rate, indicate an annual cost of excess bed capacity of $17.2 billion in 1989, $24.1 billion in 1991, and over $25 billion in 1993. Copyright 1996 by MIT Press.

Mun S. Ho, , Finding Cointegration Rank in High Dimensional Systems Using the Johansen Test: An Illustration Using Data Based Monte Carlo Simulations, The Review of Economics and Statistics, Vol. 78, No. 4 (Nov., 1996), pp. 726-732

Studies based on inputs measured at the state level generally report that school expenditures have substantial effects on students' adult wages, whereas studies based on less aggregated measures report small effects. The author uses wage data from High School and Beyond to analyze this discrepancy, and to estimate the effect of school expenditures on students' post-schooling earnings. The author finds that the discrepancy in the literature stems mostly from two factors: measurement error in district-level expenditures and omitted state effects in the earnings regression. The author also finds that the effect of school expenditures on earnings is significant but small. A 10% increase in school spending would increase students' adult wages by only 0.68%. Copyright 1996 by MIT Press.

Regressions of ex post changes in floating exchange rates on appropriate interest differentials typically imply that the high-interest rate currency tends to appreciate, the "forward discount puzzle."Using data from the European Monetary System, we find that a large part of the forward discount puzzle vanishes for regimes of fixed exchange rates.That is, deviations from uncovered interest parity appear to vary in a way which is dependent upon the exchange rate regime.By using the many EMS realignments, we are also able to quantify the "peso problem.

A parametric specification of an indirect utility function in terms of expenditure and two unit cost functions.is.proposed.Specification of these unit cost functions in terms of regular functions leads to the notion of an "effectively globally regular" system of demand equations; that is, a system of demand equations that is regular over a cone in expenditure-price space, and for which the regular region includes all points in any given sample, and all values of nominal expenditure and prices generating higher values of real expenditure than the sample minimum.This general model nests a number of popular demand systems, such as the Linear Expenditure System, as special cases.An empirical application demonstrates the value of the generalization.

We examine the evidence on whether real wages decline with age among older men. While the general human capital model of wage growth over the life cycle predicts that wages will fall as workers approach the end of their career, alternative models of wage growth do not predict these wage declines. We find that in longitudinal estimates of age-wage profiles wage declines only set in for workers in their 60s. Furthermore, these longitudinal declines are at least partly due to interactions with the Social Security system. The earnings cap or other effects of Social Security appear to lead some workers to choose jobs and job characteristics associated with lower wages. Copyright 1996 by MIT Press.