ON THE INDEPENDENCE OF THE SAMPLE MEAN AND TRANSLATION-INVARIANT STATISTICS FOR MATRIX NORMAL DISTRIBUTIONS

We present an explicit characterization of the joint dependency structure of an n×p matrix normal random matrix such that the p-dimensional sample mean vector is independent of all translation invariant statistics.     

Optimal Times to Contact Sample Households

Finding a suitable respondent at home is an essential and expensivecomponent of a household survey. This article reports on theresults of a study of the probabilities of finding someone aged14 or older at home and discusses the application of such datato survey design and budgeting.     

An evaluation of the tandem configuration automated guided vehicle system

The performance of a tandem automated guided vehicle (AGV) system is compared with that of conventional AGV track systems. In the tandem system the track is divided into non-overlapping, single-vehicle closed loops. Using simulation it is shown that, because of trips requiring delivery across loops, the tandem system has a higher expected travel time per load and thus a greater average time in system. When the loads are delivered within the original loop only, the tandem system has a lower average time in system than in conventional systems, although the difference is modest. Track layout appears to influence the average time in system much more than docs the dispatching rule invoked. The success of the tandem system is highly dependent on approximately even AGV utilization, which may be difficult to obtain in practice.     

Job queues, discrimination, and affirmative action

If employers have different Becker-type discrimination coefficients for different demographic groups, then the implementation of affirmative action may have a differential impact on those groups. We estimate two vector autoregressive models of the U.S. economy, including the unemployment rates of four demographic groups. We find that a job queue existed before the implementation of affirmative action and that affirmative action changed the ordering of the job queue in manner that had a negative impact on nonwhite males. We find evidence that affirmative action may have increased the unemployment rate of nonwhite males by increasing their duration of unemployment.     

Television Coverage of State Government

An empirical analysis of 3,334 state and local television newsstories in 10 cities clarifies the reluctance of commercialtelevision newscasters to cover state government. Contrary towhat many television news directors assert, television's minimalattention to state government cannot be attributed to an allegednegative relationship between state government and newscastratings. Rather, it appears to stem from the high cost of obtainingfilmed coverage of state government in a timely fashion.     

ABORTION AND CRIME: UNWANTED CHILDREN AND OUT-OF-WEDLOCK BIRTHS

Legalizing abortion can either increase or decrease investments in children's human capital. This article finds that abortion increases the number of out-of-wedlock births. Using data that more directly links the criminal with age when the crime was committed, not age when arrested, and fixing the assumption in previous research that no abortions took place prior to the Roe v. Wade decision in the 45 states affected by that decision, we find consistent significant evidence that legalizing abortions increased murders by over 7%. Linear estimates indicate that legalization increased total annual victimization costs by at least $3.2 billion. ( JEL K42, K14, J24)     

A STUDY OF AUTOMOBILE INVENTORY INVESTMENT

The volatile fluctuations in automobile inventory levels are explained well by the stock-adjustment model estimated in this paper on monthly time-series data. In contrast to the findings of most previous inventory studies, fluctuations in financial inventory carrying costs (as measured by a service cost of capital) are found to influence the target auto stocks in both a statistically significant and economically important manner. Also, the speed-of-adjustment of automobile inventories to their target level is estimated to be relatively fast, about thirty percent per month.