Knowledge that Transforms

In their seminal study on interindustry wage differentials, Krueger and Summers (1988) expressed estimated industry differences as deviations from a hypothetical employment-share weighted mean. Virtually the whole labor literature has followed their approach, yet most studies avoid calculating the exact standard errors of these differences. This note relates this problem to the general literature on dummy variables and their interpretation. It is demonstrated that the implementation of exact estimates involves only simple matrix operations, making any approximative procedure difficult to justify. Disregarding this conclusion will in practice, even with large samples, lead to substantially overstated standard errors of the estimated differentials and to the understatement of their overall variability.

Equality and efficiency are key issues in educational reform. Here the authors analyze the efficiency and equality consequences of various school finance reforms using a cost-indirect output distance function. This function readily models multiple-output production under conditions of budgetary constraint, and provides a natural measure of performance that is closely related to Farrell-type measures of efficiency. The analysis suggests that despite school district inefficiency, finance reforms can affect student achievement. However, any potential gains in output from redistribution are dwarfed by the potential gains from increased efficiency. More strikingly, the analysis demonstrates that budgetary reforms designed to equalize expenditures could actually increase the inequality of student achievement.

This paper studies exchange rate pass-through in U.S. manufacturing industries and its cross-sectional variation. Through an adapted Dixit–Stiglitz model of product differentiation, the paper predicts that pass-through is positively related to the degree of product differentiation and inversely related to the elasticity of marginal cost with respect to output. Empirical estimates of the pass-through elasticities show that pass-through is incomplete and varies across industries. The degree of pass-through is found to be positively correlated to different proxies for product differentiation, and negatively to a proxy for the elasticity of marginal cost.

We assess the cost-reducing impacts of increasing stocks of “high-tech” equipment (O capital). Our empirical analysis is based on a dynamic production theory model and annual data for two-digit U.S. manufacturing industries (1952–1991). We find evidence of overinvestment in O capital in the mid to late 1980s, following a period of strong investment incentives in the late 1970s. By the end of the 1980s, however, the returns to investment and falling prices for O capital more than justified the high investment levels in nondurable-goods industries, and the benefit–cost ratio was also increasing for durable-goods industries. The underlying substitution patterns suggest that high-tech capital expansion increases demand for most capital and noncapital inputs overall, but saves on materials inputs. In durables industries, however, both energy and “other” capital appear somewhat substitutable with O capital, and in nondurables industries increasing high-tech intensity may be a factor underlying stagnating labor demand.“We see computers everywhere except in the productivity statistics.”Attributed to Robert M. Solow

We use disaggregated data on trade flows, production, and trade barriers for 41 countries in 1988 to examine the political and economic determinants of non-tariff barriers, as well as the impact of protection (both tariff and non-tariff) on trade flows. We use an econometric framework that allows for the simultaneous detennination of trade barriers and trade flows. Our results are consistent with political-economy theories of the determinants of protection: even after accounting for industry-specific factors, nations tend to protect industries that are weak, in decline, and threatened by import competition. Countries also give more protection to large industries; these might be thought of as politically important. Nations use tariffs, non-tariff barriers, and exchange rate controls as complementary instruments of protection.

We analyze the impact of increased immigration on employment outcomes of natives in Germany using a data set of county-level variables for the late 1980s. In order to construct more unified labor market regions, we aggregate the 328 counties to 167 larger regions. We study two measures of immigration, the change in the share of foreigners between 1985 and 1989 as well as one-year gross and net flows of immigrants to an area. In order to address the potential problem of immigrant selection into local labor markets, we condition on previous labor market outcomes, which may serve as the basis of immigrant selection. This specification allows for mean reversion in the unemployment rate, which is strong in our data set and period of study. We show that this rules out some other approaches of identifying the impact of immigration. Our results indicate no detrimental effect of immigration. We find no support for the hypothesis that the absence of displacement effects is due to a response of native migration patterns.

We take a model selection approach to the question of whether a class of adaptive prediction models (artificial neural networks) is useful for predicting future values of nine macroeconomic variables. We use a variety of out-of-sample forecast-based model selection criteria, including forecast error measures and forecast direction accuracy. Ex ante or real-time forecasting results based on rolling window prediction methods indicate that multivariate adaptive linear vector autoregression models often outperform a variety of (1) adaptive and nonadaptive univariate models, (2) nonadaptive multivariate models, (3) adaptive nonlinear models, and (4) professionally available survey predictions. Further, model selection based on the in-sample Schwarz information criterion apparently fails to offer a convenient shortcut to true out-of-sample performance measures.

As with most analyses involving microdata, applications of count data models must somehow account for unobserved heterogeneity. The count model literature has generally assumed that unobservables and observed covariates are statistically independent. Yet for many applications this independence assumption is clearly tenuous. When the unobservables are omitted variables correlated with included regressors, standard estimation methods will generally be inconsistent. Though alternative consistent estimators may exist in special circumstances, it is suggested here that a nonlinear instrumental-variable strategy offers a reasonably general solution to such estimation problems. This approach is applied in two examples that focus on cigarette smoking behavior.

This paper studies the least squares estimation of a change point in multiple regressions. Consistency, rate of convergence, and asymptotic distributions are obtained. The model allows for lagged dependent variables and trending regressors. The error process can be dependent and heteroskedastic. For nonstationary regressors or disturbances, the asymptotic distribution is shown to be skewed. The analytical density function and the cumulative distribution function for the general skewed distribution are derived. The analysis applies to both pure and partial changes. The method is used to analyze the response of market interest rates to discount rate changes.

Ever since Nelson and Plosser (1982) found evidence in favor of the unit-root hypothesis for 13 long-term annual macro series, observed unit - root behavior has been equated with persistence in the economy. Perron (1989) questioned this interpretation, arguing instead that the "observed" behavior may indicate failure to account for structural change. Zivot and Andrews (1992) restored confidence in the unit-root hypothesis by incorporating an endogenous break point into the specification. By allowing for the possibility of two endogenous break points, we find more evidence against the unit-root hypothesis than Zivot and Andrews, but less than Perron.