Knowledge that Transforms

A survey of several economic journals reveals that very often the Durbin-Watson and the portmanteau (Box-Pierce or Ljung-Box) tests are inappropriately applied to linear models with lagged dependent variables and exogenous regressors. Sampling experiments indicate that the Durbin-Watson performs poorly in models with more than one lag of the dependent variable, a situation commonly considered in the literature. The experiments also indicate that the portmanteau test is inadequate when applied to dynamic linear models with exogenous regressors. In addition, the performance of Durbin's h and m tests in models commonly used in the literature but not considered by previous studies is evaluated. The results reveal that among the four tests examined, the one which is the least frequently used in practice (the m test) has the best performance.

Joseph V. Terza, Paul W. Wilson, Analyzing Frequencies of Several Types of Events: A Mixed Multinomial- Poisson Approach, The Review of Economics and Statistics, Vol. 72, No. 1 (Feb., 1990), pp. 108-115

A presidential vote function and a presidential approval ratings function are jointly estimated for U.S. post-war observations. The estimation technique treats the two equations as seemingly unrelated regressions with unequal numbers of observations. Cross-equation restrictions implying that voters and poll respondents use identical standards in judging the economic performance of incumbents are imposed and tested. Estimates show that both votes and approval ratings are influenced by GNP growth and inflation. The results suggest that poll respondents are more inflation averse than voters; however, tests of this hypothesis are not conclusive.

This paper investigates why female employees are overrepresented in the private nonprofit sector. Using the 1977 Quality of Employment Survey, the author determines the importance of occupational locus and compensation structure as explanations for the large number of women in the sector. While occupational locus does contribute to the large number of women in the sector, the structure of nonprofit compensation is also an important factor in drawing women to the sector. Controlling for differences in compensation patterns of for-profit and nonprofit firms, women are not more likely than men to work in the nonprofit sector.

We discuss the problem of estimating M (>1) high-frequency (say, quarterly or monthly) time series using the relevant low-frequency (say, annual or quarterly) data, the sum for each intra-annual period of the series to be estimated and, finally, a number of related indicators. The optimal (in least squares sense) estimator that fulfills both temporal and contemporaneous aggregation constraints is derived. In addition, we critically comment on the estimation approach followed by Rossi (1982), showing that a convenient reformulation of that method can be viewed as a special application of a known adjustment technique.

B. Kelly Eakin, Daniel P. McMillen, Mark J. Buono, Constructing Confidence Intervals Using the Bootstrap: An Application to a Multi-Product Cost Function, The Review of Economics and Statistics, Vol. 72, No. 2 (May, 1990), pp. 339-344

Non-cooperative bidding theory for sealed-bid auctions generally implies testable predictions that are conditioned on the risk attitudes of agents. Received laboratory experiments that purport to test this theory do not generally control for the risk attitudes of subjects. Those experiments exhibit behavior inconsistent with popular bidding models that assume that agents have the same aversion to risk or are all risk neutral. This paper constructs an explicit Bayesian prior distribution for the risk attitudes of experimental subjects and reconsiders the experimental results. It finds that observed bidding behavior is still inconsistent with the Nash predictions when explicit prior weights are attached to alternative assumptions about subject risk attitudes. Thus one cannot account for observed bidding anomalies by appealing to uncontrolled nuisance variables such as risk attitudes. Non-cooperative bidding theory for sealed-bid auctions generally implies testable predictions that are conditioned on the risk attitudes of agents. Archetypical of this result is the Nash Equilibrium prediction for First Price auctions for an object that is valued by agents in an independent and private manner. Received laboratory experiments that purport to test this theory do not generally control for the risk attitudes of subjects. Those experiments exhibit behavior inconsistent with popular bidding models that assume that agents have the same aversion to riskor are all risk neutral. In this paper we construct an explicit prior distrlbution for the risk attitudes of experimental subjects and reconsider the experimental results. We find that observed bidding behavior is indeed consistent with the Nash predictions when explicit prior weights are attached to alternative assumptions about subject risk aversion. However, when one allows for risk loving subjects as well, observed behavior is inconsistent with Nash predictions. Thus one cannot account for observed bidding anomalies by appealing to uncontrolled nuisance variables such as risk attitudes. In section I we consider a specific Nash Equilibrium (NE) bidding model due to Cox, Roberson and Smith (1982) and Cox, Smith and Walker (1988) that clearly illustrates the risk-sensitivity of the theoretical predictions. In section II we provide independent evidence of the risk attitudes of experimental subjects in a test for risk attitudes developed by Harrison (1986a). This evidence allows us to construct an explicit prior probability density function over the coefficient of (constant relative) risk attitudes employed in the specific bidding model of section I. In section III we reconsider the evidence from the First Price (FP) experiments reported in Cox, Roberson and Smith (1982) and Cox, Smith and Walker (1983a, 1983b). I. A Specific Bidding Model Cox, Roberson and Smith (1982), hereafter CRS, present a model based on a power function utility specification for agent i: Ui(y) y= (1) Received for publication August 31, 1987. Revision accepted for publication December 18, 1989. * University of South Carolina. I am grateful to two anonymous referees for helpful comments, although they are not responsible for my conclusions.

When estimating regression models of educational achievement with pre- and posttest data, researchers have overlooked a sample selection bias that may occur even where initial assignment to the control and experimental groups is random. The bias arises because students who take the pretest but do not take the posttest are excluded from the regression analysis. Using data from a nationally normed test of high school student knowledge of economics, adjustment for this bias is shown to influence the estimated effectiveness of programs aimed at increasing student learning of economics.

A model of neighborhood turnover drawn from Bond and Coulson (1989) is proposed. The type of turnover process that is obtained is shown to depend mainly on the hedonic bid functions for housing and neighborhood quality. A demand system of four hedonic attributes is estimated. The main results are that the traditional model of filtering by age of unit does not occur and that filtering by housing size does. Tipping due to changes in median neighborhood income is also quite possible. Tipping through changes in racial composition appears less likely.

This paper deals with the problems of partial equlibrium analysis by presenting estimates from a static ten sector computable general equilibrium (CGE) model of the U.S. economy calibrated to the year 1984. Following the introduction, the paper is organized as follows. Section 2 outlines the model. Section 3 details the sources of estimates of premia on preexisting QRs (quota rents) in 1984 and the sources for the parameters describing demand and supply elasticities. Welfare and employment estimates of QR removal are presented by industry and in the aggregate in Section 4. Conclusions follow in Section 5.