AN ADAPTIVE MULTIVARIATE APPROACH TO TIME SERIES FORECASTING

In recent years, time series analysts have shifted their interest from univariate to multivariate forecasting approaches. Among them, the Box-Jenkins transfer function process and the state space method have received the most attention. This paper presents a simplified approach that embodies some desirable features of existing methods. It stresses empirical analysis, has a unified modeling structure, is easily applicable, and is adaptive to changes without necessitating prior information on the evolution of a system under study. The core of the method relies on the Carbone-Longini adaptive estimation procedure (AEP). Results of a comparative study based on the well-known Lydia E. Pinkham data and the Box-Jenkins sales/leading indicator data illustrate the merits of multivariate AEP in improving forecasting accuracy while simplifying the analysis process. Subject Area: Forecasting.     

CYCLICAL BEHAVIOR OF UNEMPLOYMENT AND THE HELP WANTED INDEX: A CROSS SPECTRAL ANALYSIS

The unemployment rate and the help-wanted index are two frequently used indicators of the state of the labor market. Traditional regression techniques used in the analysis of the relationship between these two indicators and their cyclical behavior with business fluctuations have yielded varying results. Spectral analysis is used in this article to examine the cyclical behavior of the labor market indicators with respect to each other and with respect to a measure of aggregate economic activity in the United States. Very strong relationships are found, lending support to the rationale of using such indicators as labor market proxies. This application of spectral analysis also provides an illustration of the potential fruitfulness of spectral analysis in examining the cyclical relationships between economic time series.     

DETECTION OF TURNING POINTS IN A TIME SERIES

Consider a time series for which, over a finite interval, there is a model that provides an adequate forecast of the series. Without loss of genality one can take the series over this period of time to have mean O and variance s̀². At some time the underlying process changes to some other model. The previous model no longer produces errors with mean zero. It is assumed, however, that the variance remains s̀². The problem considered here is to detect the change in the process as quickly as possible after it happens. The technique is a computationally feasible extension of Wald's [14] sequential analysis, to develop a parabolic mask centered over the most recent cumulative sum (“cusum”) of the forecast errors. Detection occurs when any previous point in the series of cusums lies outside the parabola. The technique is illustrated by an APL program applied to the logarithms of weekly changes in closing prices for IBM common stock on the New York Stock Exchange over the period 1968–1970.     

ACCURACY OF JUDGMENTAL FORECASTING OF TIME SERIES

Experiments contrasted judgmental and objective forecast methods. Judgmental methods included “eyeball” extrapolation of time-series plots and judgmental adjustment. Objective methods included Box-Jenkins (BJ), Carbone-Longini AEP filtering (CL), Holt-Winters (HW), and other smoothing techniques. Objective methods proved more accurate than eyeball extrapolation. However, judgmental adjustment improved the accuracy of some objective forecasts.     

A NOTE ON A MULTIVARIATE GENERALIZATION OF THE KRUSKAL-WALLIS TEST

Marketers are often interested in testing whether the mean vectors of multivariate distributions are equal. The test usually applied, one-way MANOVA, assumes the distributions are multinormal. Unfortunately, this assumption is not supported in many studies. As an alternative, a nonparametric multivariate one-way analysis of variance procedure is presented.     

THE LOT OF THE UNEMPLOYED: A TIME USE PERSPECTIVE

This paper provides new evidence on the time use of employed and unemployed individuals in 14 countries. We devote particular attention to characterizing and modeling job search intensity, measured by the amount of time devoted to searching for a new job. Job search intensity varies considerably across countries, and is higher in countries that have higher wage dispersion. We also examine the relationship between unemployment benefits and job search.     

A SPECTRAL ANALYSIS OF THE INTERACTION BETWEEN INVENTORIES AND SALES OF MERCHANT WHOLESALERS

In this study cross-spectral methods are used to test the hypothesis that inventory investment leads sales of merchant wholesalers through the short-term, seasonal frequencies, but sales lead inventories through the long-term, cyclical frequencies. An intuitive explanation and a time-domain, generating model for the reversal phenomenon are given.     

THE INEFFICIENCY OF WORKER TIME USE

Much work is carried out in short, interrupted segments. This phenomenon, which we label task juggling, has been overlooked by economists. We study the work schedules of some judges in Italy documenting that they do juggle tasks and that juggling causally lowers their productivity substantially. To measure the size of this effect, we show that although all these judges receive the same workload, those who juggle more trials at once instead of working sequentially on few of them at each unit of time, take longer to complete their portfolios of cases. Task juggling seems to have no adverse effect on the quality of the judges' decisions, as measured by the percent of decisions appealed. To identify these causal effects we estimate models with judge fixed effects and we exploit the lottery assigning cases to judges. We discuss whether task juggling can be viewed as inefficient, and provide a back‐of‐the‐envelope calculation of the social cost of longer trials due to task juggling.     

STRUCTURING THE LEISURE MARKET—A MULTIVARIATE ANALYSIS

The derivation of an empirically based classification scheme for the leisure market is discussed from both a substantive and methodological viewpoint. The substantive findings relate to the manner in which respondents structured relationships among 12 leisure activities and to the criteria they used in doing so. Based on the apparent structure, a number of activity types were identified. The methodological thrust of the article provides one example of how a variety of multivariate methods (multidimensional scaling, property fitting, cluster analysis and factor analysis) can be employed to form a system of complementary techniques aimed at providing an in-depth understanding of data and the phenomena they describe.     

PERFORMANCE DIFFERENCES IN THE USE OF GRAPHIC AND TABULAR DISPLAYS OF MULTIVARIATE DATA

It has been suggested that graphically displayed multivariate data help decision makers better understand information they are called on to analyze. This study compares judgments made from one recently suggested multivariate display technique with judgments made from traditional tabular displas of financial figures. Significant differences in task performance are found to be related both to differences in the stimulus sets and to individual differences among the subjects. Our results suggest that the relative contribution of graphic displays to decision making may vary considerably from situation to situation.     

THE ANALYSIS OF OUTLYING DATA POINTS USING ROBUST REGRESSION: A MULTIVARIATE PROBLEM-BANK IDENTIFICATION MODEL*

Because the eight largest bank failures in United States history have occurred since 1973 [24], the development of early-warning problem-bank identification models is an important undertaking. It has been shown previously [3] [5] that M-estimator robust regression provides such a model. The present paper develops a similar model for the multivariate case using both a robustified Mahalanobis distance analysis [21] and principal components analysis [10]. In addition to providing a successful presumptive problem-bank identification model, combining the use of the M-estimator robust regression procedure and the robust Mahalanobis distance procedure with principal components analysis is also demonstrated to be a general method of outlier detection. The results from using these procedures are compared to some previously suggested procedures, and general conclusions are drawn.     

ON PREDICTING COMPUTATIONAL TIME OF A BRANCH AND BOUND ALGORITHM FOR THE ASSIGNMENT OF FACILITIES

Experience with branch and bound algorithms indicates that computational time is a function of not only the size of the problem, but also the nature of the input data. This paper formulates statistically-based variables which describe certain characteristics of the input data and experimentally evaluates their ability to predict computational time for one branch and bound algorithm, the relative location of facilities or “plant layout” problem. Results suggest that the described experimental procedure may be useful for an a priori assessment of the computational difficulty of specific branch and bound problems.     

DUAL-METRIC MEASUREMENT OF THE IMPACT OF COMPLEXITY ON DECISION QUALITY*

Both sophisticated and unsophisticated subjects were provided either a simple or a complex cue set at random on which to base decisions. A dual-metric approach that consisted of measuring interrater consistency and cross-item variance for each subject category was used to compare the decision quality of the subjects. Although complexity and decision quality were inversely related, the sophisticated subjects performed relatively better using the complex cue set; the unsophisticated subjects performed better using the simple cue set. These findings support the notion that decision-maker and task compatibility are important to the assessment of decision quality.     

THE MEASUREMENT AND OPERATING BENEFITS OF COMPONENT PART COMMONALITY

The product structure is a key input to a material requirements planning system. Yet, the effect of alternate product structure designs on system performance is not well understood. Analytical measures are needed to accurately relate product-structure characteristics to total cost and delivery performance. A new analytical measure of product structure called the degree of commonality index is defined in this paper. The degree of commonality index can be applied to single end items, product families, entire product lines, or to any level of product structure. This study suggests that component part commonality, as measured by the degree of commonality index, has a significant effect on system performance. For example, company standardization programs that increase the degree of component part commonality result in manufacturing cost reductions. The degree of commonality index provides a way to directly relate the degree of component part commonality to various dependent variables such as total cost, work center load, and delivery performance.     

AN INDEX FOR THE MEASUREMENT OF CONSUMER LOYALTY*

The concept of consumer “loyalty” (i.e., preference for a particular store or brand) is intriguing to marketing scholars and practitioners. Research in this area would be enhanced by the acceptance of an operational measure of consumer loyalty. Loyalty appears to be a tri-dimensional concept; per cent of budget, allocated to the store or brand, amount of switching, and number of alternatives explored have been used to measure loyalty. Consequently, the authors propose a “loyalty index” which combines these three measures. An empirical test of the proposed index showed that the index did discriminate among degrees of consumer loyalty, and was related to consumer psychological characteristics in a manner similar to that of the generally used single measures of consumer loyalty.     

A RETROSPECTIVE STUDY ON THE INFLUENCE OF A STATE PARK-LAKE ON LAND VALUE,FROM THE TIME OF LAND ACQUISITION TO RESERVOIR FILLING

A new multiple purpose park-lake facility is being developed in the sub urbanrural fringe of the Philadelphia, Pennsylvania metropolitan region. This study assesses the impact of the new park-lake upon surrounding land values in order to provide information for improved planning and decision making in such a context. The investigation covers the period from the time of land acquisition to that of near-completion of dam construction. The results of the study provide a “baseline” from which it will later be possible to assess changes in land values as the park-lake becomes fully operational as a major public recreational facility. Two geographic zones of impact on surrounding land value are identified by the analysis, and variations in the impact of the park-lake over time are identified. The importance of these new findings to public organizations involved in recreational decision making is discussed and areas for future research are suggested.     

THE HERMITE DISTRIBUTION AS A MODEL OF DEMAND DURING LEAD TIME FOR SLOW-MOVING ITEMS

Heretofore, the Poisson and the Laplace distributions have been used to model demand during lead time for slow-moving items. In this paper, we present a Poisson-like distribution called the Hermite. The advantage of the Hermite is that it is as simple to use as the Poisson and the Laplace are. Moreover, the Hermite is the exact distribution of demand during lead time when unit demand is Poisson, P(Λ), and lead time is normally distributed, N(μ, σ²), so long as (μ/σ²)≥Λ. Thus, the Hermite can enhance the accuracy of analysis as well as add to the tools available to the analyst.