Industrial Location Modeling: Extending the Random Utility Framework*

Abstract. Given sound theoretical underpinnings, the random utility maximization‐based conditional logit model (CLM) serves as the principal method for applied research on industrial location decisions. Studies that implement this methodology, however, confront several problems, notably the disadvantages of the underlying Independence of Irrelevant Alternatives (IIA) assumption. This paper shows that by taking advantage of an equivalent relation between the CLM and Poisson regression likelihood functions one can more effectively control for the potential IIA violation in complex choice scenarios where the decision maker confronts a large number of narrowly defined spatial alternatives. As demonstrated here our approach to the IIA problem is compliant with the random utility (profit) maximization framework.     

Exposing Unobserved Spatial Similarity: Evidence from German School Choice Data

In a spatial context, flexible substitution patterns play an important role when modeling individual choice behavior. Issues of correlation may arise if two or more alternatives of a selected choice set share characteristics that cannot be observed by a modeler. Multivariate extreme value (MEV) models provide the possibility to relax the property of constant substitution imposed by the multinomial logit (MNL) model through its independence of irrelevant alternatives (IIA) property. Existing approaches in school network planning often do not account for substitution patterns, nor do they take free school choice into consideration. In this article, we briefly operationalize a closed‐form discrete choice model (generalized nested logit [GNL] model) from utility maximization to account for spatial correlation. Moreover, we show that very simple and restrictive models are usually not adequate in a spatial choice context. In contrast, the GNL is still computationally convenient and obtains a very flexible structure of substitution patterns among choice alternatives. Roughly speaking, this flexibility is achieved by allocating alternatives that are located close to each other into nests. A given alternative may belong to several nests. Therefore, we specify a more general discrete choice model. Furthermore, the data and the model specification for the school choice problem are presented. The analysis of free school choice in the city of Dresden, Germany, confirms the influence of most of the exogenous variables reported in the literature. The estimation results generally indicate the applicability of MEV models in a spatial context and the importance of spatial correlation in school choice modeling. Therefore, we suggest the use of more flexible and complex models than standard logit models in particular. En un contexto espacial, los patrones sustitución flexible juegan un papel importante en el modelamiento del comportamiento de las decisiones individuales. Varios problemas de correlación pueden presentarse si dos o más alternativas de elección comparten características no observables por el modelador. Los modelos de valor extremo (multivariate extreme value‐MEV) ofrecen la posibilidad de relajar la propiedad de sustitución constante (constant substitution) presente en los modelos logit multinomiales (multinomial logit‐MNL), a través de su propiedad de independencia de alternativas irrelevantes (Independence of irrelevant alternatives property ‐IIA). A menudo, los enfoques existentes en la planificación de redes escolares no toman en consideración los patrones de sustitución y de libre elección de escuela. En este artículo, los autores presentan brevemente el funcionamiento de un modelo de elección discreta (discrete choice model) para la maximización de utilidad o modelo logit anidado generalizado (generalized nested logit model‐GNL) para dar cuenta de la autocorrelación espacial. Los autores sostienen que modelos demasiado simples y restrictivos no suelen ser adecuados en un contexto de elección espacial. En contraste el modelo GNL es conveniente en términos de su computación y obtiene una estructura muy flexible de los patrones de sustitución entre las alternativas de elección. En términos generales, esta flexibilidad se logra mediante la asignación (o anidación) de las alternativas cercanas en el espacio (una alternativa puede pertenecer a varios nidos). Por lo tanto, los autores presentan un modelo de elección discreta más general. El estudio presenta además datos y la especificación del modelo para un caso de elección de escuela concreto: el análisis de libre elección de escuela en la ciudad de Dresden, Alemania. El análisis confirma la influencia de la mayoría de las variables exógenas presentes en la literatura. Los resultados de la estimación demuestran en términos generales la aplicabilidad de los modelos MEV en un contexto espacial y la importancia de la autocorrelación espacial en el modelado de elección de escuela. Los autores concluyen sugiriendo el uso de modelos más flexibles y complejos que los modelos utilizados habitualmente, en particular los modelos logit estándar. 从空间视角看,灵活的替代模式在个人行为选择建模中发挥着重要作用。当存在两个或两个以上备选方案集具有共性且无法被建模者观察到时,就可能出现相关性问题。多元极值模型（MEV）通过不相关的替代属性（IIA）实现了对多元logit模型（MNL）中常数限制的松弛替代。现有校园网络规划方法通常无法解释替代模式,而且没有考虑到自由择校因素。本文简要地建立一个封闭离散选择模型（广义嵌套（GNL）模型）,从效用最大化角度来解释空间相关性。此外分析还表明,非常简单的约束模型通常不具有足够的空间选择情境。相比之下,GNL模型计算便捷,且可以在各选择方案中获得非常灵活的替代模式。大致而言,这种灵活性大体是通过与住处位置距离上彼此靠近的替代选择分配而获得,一个给定的选择可能属于不同的住处。因此,我们给出了一个更一般的离散选择模型。此外,还给出了针对择校问题的数据和模型设定。基于德国德累斯顿市自由择校分析,证实了已有研究中多数外生变量的影响。估计结果证实了MEV模型在空间分析中的适用性以及择校模型中空间相关的重要性,并建议使用更加灵活和复杂的模型而不是标准的logit模型。     

Studying Rules Changes in the U.S. House: Evidence from an Alternative Empirical Model

Binder (1996) and Schickler (2000) define the current debate as to why the U.S. House has changed its standing rules regarding the majority rule and the minority rights. I revisit their empirical models—binary logit and ordered logit—and theoretically and statistically test the appropriateness of these models. I find that both of them are actually choosing inappropriate models. Their theoretical claims cannot be properly examined by utilizing their choices of models. In addition, the data do not satisfy the “parallel regression” assumption but do satisfy the “independence of irrelevant alternatives” assumption, which supports using an alternative multinomial logit model. I further extend the model, and find the dynamic nature of rules changes in the U.S. House. It appears there is no symmetry between the rules changes that promote the majority rule and the rules changes that enhance the minority rights.     

Estimating a Destination-Choice Model from a Choice-based Sample with Limited Information

In most applications of multinomial logit and other probabilistic discrete-choice models, the estimation data set is either a simple random sample of the population of interest or an exogenously stratified sample. Often, however, it is cheaper and easier to sample individuals while they are carrying out the chosen activity of concern. This produces a choice-based sample, which presents important problems of estimation and inference. This paper is concerned with estimation of destination-choice models from choice-based samples when neither the aggregate market shares of alternatives nor the probability distribution of explanatory variables in the population is known. The method of Cosslett (1981) for estimating multinomial logit models from such data is summarized, and the limitations on information about choice behavior that can be recovered from the sample are explained. An empirical model of pharmacy choice in the Namur, Belgium, area is presented. It is shown that useful and important information about destination-choice behavior can be obtained from a choice-based sample, even without knowledge of aggregate market shares and the probability distribution of explanatory variables.     

HOUSEHOLD LOCATION AND RACE: A 20‐YEAR RETROSPECTIVE*

ABSTRACT In research conducted some 20 years ago, we elucidated the starkly lower suburbanization propensities of black households in the Washington, D.C. metropolitan area. The paper showed that simulated closure of large socio‐economic gaps between blacks and whites did little to diminish prevailing high levels of residential segregation or otherwise enhance moves by black households to areas of educational, employment, and housing opportunity. Some two decades later and in the wake of significant urban evolution, this paper assesses anew racial variations in residential location choice. Results of the multinomial logit (MNL) analysis indicate large, persistent racial differentials in intrametropolitan residential location choice. While black location choice in 2000 was relatively more dispersed than in 1980, it remained remarkably concentrated in D.C. and Prince George's County. As in our prior analysis, results showed that large simulated gains in black economic and educational status had little effect on prevailing racial segregation. These findings underscore the ongoing, limited access of black households to schooling, employment, and homeownership opportunities available outside traditional areas of settlement. In marked contrast, the locational choices of Latino and immigrant households bore greater similarity to those of whites and were more sensitive to improvements in socio‐economic status.     

Preferences for Housing, Jobs, and Commuting: A Mixed Logit Analysis

This paper reports stated preferences of Dutch workers for combinations of housing, employment, and commuting. The analysis uses standard logit models as well as mixed logit models. Estimation results offer insights into the relative importance of various aspects of housing, employment, and commuting. Households dislike commuting and the value of commuting time implied by the model is high in comparison to the wage rate. Nevertheless, preferences for some housing attributes are strong enough to make substantially longer commuting acceptable to most workers. Of special interest is the strong preference for living in small-or medium-size cities, especially among two income households. Using a mixed logit model instead of a standard logit model results in a substantial improvement of the loglikelihood, reflecting the importance of heterogeneity among respondents. If no individual characteristics are incorporated into the model, the mixed logit implies substantially lower average monetary evaluations of most attributes. These differences are much smaller if some individual characteristics are incorporated into the model.     

Strategic voting and constituency context: Modelling party preference and vote in multiparty elections

We examine the role of strategic motivations in mediating the relationship between underlying political preferences and vote choice, in a multiparty, single member, simple plurality system, and examine the role of constituency context in determining the scope for strategic voting. Political preference data from the British Election Panel Survey, 1997–2001, were modelled with mixed multinomial logit models. Latent variables were used to model the stable party political traits underlying observed preferences, allowing correlation between choices and so avoiding the restrictive assumption of independence from irrelevant alternatives. Ranked approval ratings were used to characterize the underlying political preferences in the presence of insincere voting. From these models we estimate that approximately 9% of votes cast may have been affected by strategic factors. In keeping with ‘Duvergers law’, the smallest of the three main parties, the Liberal Democrats, were found to be most affected by strategic voting.     

DISCRETE SPATIAL CHOICE MODELS FOR AGGREGATE DESTINATIONS*

ABSTRACT. In problems of spatial choice, the choice set is typically more aggregated than the one considered by decision-makers, often because choice data are available only at the aggregate level. These aggregate choice units will exhibit heterogeneity in utility and in size. To be consistent with utility maximization, a choice model must estimate choice probabilities on the basis of the maximum utility within heterogeneous aggregates. The ordinary multinomial logit model applied to aggregate choice units fails this criterion as it is estimated on the basis of average utility. In this paper, we derive and discuss a model which utilizes the theory underlying the nested logit model to estimate the appropriate maximum utilities of aggregates. We also demonstrate that the aggregate alternative error terms are asymptotically Gumbel, thereby relaxing the assumption of extreme value distributed error terms. This is accomplished with help from the asymptotic theory of extremes.     

Modeling Strategies for Categorical Data: Examples from Housing and Tenure Choice

Models to investigate categorical data can be divided into preprocessing, limited parameterization, and formal logit models. To illustrate the advantages of preprocessing and limited parameterization models they are applied to a data set of tenure and type of housing choice before the data are examined with hierarchical logit and nested logit models. The preprocessing approaches are useful in selecting optimal subsets of independent variables with respect to the dependent variable. The ease of application and interpretation of a limited parameterization approach extends the clarity of the results from the preprocessing approaches. Because some variables are only relevant at specific levels of other independent variables, nonstandard (nested) logit models are necessary to understand the nested relationships.     

LEAST SQUARES ESTIMATES OF THE REGIONAL EMPLOYMENT MULTIPLIER: AN APPRAISAL*

Basically, we have attempted to show the following in the course of setting out the algebra of regression analysis of selected regional employment multiplier models: (1) When the basic features of the model are shaped by the assumption of an unlagged response of local employment to changes in export employment, the least squares estimates of the multipliers are highly sensitive to the export coefficients vector A, given the sample observation matrix X. In a completely disaggregated model such as Equation (21), the multipliers are solely determined by the export coefficients and thus are entirely independent of sample observations. However, this independence does not hold in the case of a partially disaggregated model. The identity relation is also destroyed when a lag relationship is introduced into a completely disaggregated model. (2) A simple lag model produces results bascially different from those obtained by an unlagged model if the overall differences between current and lagged observations are significant. (3) Given a matrix of sample observations on employment, it is possible to estimate the upper limits of a least-squares aggregate multiplier and its variance simply from knowledge of the export coefficients (4) The export coefficients vector has also an important bearing upon the correlation coefficient. The correlation is unity if and only if the export coefficients vector is proportional to the local employment coefficients vector, while it is zero if and only if the export coefficients vector is a vector all of whose elements are one. Also, the correlation coefficient is equal to one when a completely disaggregated model is used. There is finally the question of what these results mean in terms of the formulation of a multiplier model. First of all, in view of the crucial importance of the export coefficients and the difficulties of estimating them, most of the existing models do not seem to offer promising results. Furthermore, all the models examined here have made some simplifying assumption with respect to the constancy of the export coefficients. It remains highly uncertain whether these coefficients are reasonably stable over time. Of course, it would be theoretically more acceptable to relax the assumption of the invariance of export coefficients and to obtain such coefficients at different points of time for each industry. However, this would be accomplished only at the cost of increased difficulties of estimating larger numbers of export coefficients. In addition, there is some doubt as to the validity of the assumption that export employment is proportional to export sales. Since a lag relationship is important not only in terms of attempts to formulate multiplier models more realistically, but also in terms of its significant effect on the multiplier values obtained, the nature and the form of a lagged response and its estimation problems need to be investigated in depth. Finally, problems of least squares bias and efficiency, inference, and prediction which may arise in the context of various models presented here remain to be investigated. A detailed analysis of such problems must be the subject of further investigation.     

MIGRATION AS A WAY TO DIVERSIFY: EVIDENCE FROM RURAL TO URBAN MIGRATION IN THE U.S.

This paper extends the utility maximization model of migration by introducing income and unemployment‐related uncertainties as determinants of utility, and analyzes the effects of the informational advantages of migrants. The paper maintains that migration would expand an individual's economic choices and opportunities and allow diversification. Consequently, diversification advantages influence the location decisions of migrants, an effect captured by the correlation of incomes at the origin and potential destinations. We use the discrete choice model based on random utility maximization as the framework for our empirical investigation of migration from the United States rural to urban counties. This paper takes advantage of an equivalent relation between the conditional logit model and Poisson regression to study the migration decisions using aggregate data among a large set of spatial alternatives. The results show that the diversification concerns have significant effects on location decisions of the rural‐urban migrants in the United States.     

RETURN AND ONWARD MIGRATIONS IN CANADA, 1976-1981: AN EXPLANATION BASED ON PERSONAL AND ECOLOGICAL VARIABLES

This paper applies a three-level nested logit model to the micro data of the 1981 Canadian census to explain the 1976-81 interprovincial migration choices of the non-natives (those whose province of residence was different from province of birth), aged 20 to 44, by personal factors and provincial attributes. Important personal factors include mother tongue, level of education, family type, and age. Influential provincial attributes include economic variables (income level, employment growth, and unemployment), distance, and cultural similarity. The main finding is that not only onward migrants but also return migrants were sensitive to the interprovincial variation in economic opportunities.
Dans cet article, nous appliquons un modèle logistique embotéà trois niveaux à des données micro du recensement canadien de 1981 afin ďexpliquer les choix effectués en matière de migration interprovinciale par les non-natifs (ceux résidant hors de leur province de naissance) gés de 20 à 44 ans). Les variables explicatives mises en valeur incluent des variables personnel les (langue maternelle, niveau ďinstruction, genre de famille et ge), des variables écologiques essentiellement économiques (niveau de revenu, croissance de Ľemploi et chdmage) ainsi que des variables relationnelles (distance et similarity culturellé). Selon le principal résultat obtenu, non seulement les migrations de depart mais aussi les migrations de retour sont sensibles aux écarts interprovinciaux en termes ďoccasions économiques.     

Exogenous Workplace Choice in Residential Location Models: Is the Assumption Valid?

This paper examines the assumption implicit in most models of residential location that the choice of workplace is exogenously determined. Monocentric models have generally made this assumption, but it has come under increasing scrutiny. A nested logit model of workers' choices of workplace, residence, and housing tenure within the Dallas-Fort Worth metropolitan area is developed to provide a test. A unique dataset that includes the workplace and residence census tracts of workers in Dallas-Fort Worth, and their demographic and socioeconomic characteristics, is used to estimate the model. The results confirm that a joint choice specification better represents actual choice behavior in a multinodal metropolis.     

ON FRONTIER SPECIFICATIONS FOR DISCRETE BINARY CHOICE ANALYSIS

ABSTRACT. In estimating a discrete choice model one is actually estimating the parameters of a conditional indirect utility function. I explore the consequences of recognizing that this function is a maximum-value (frontier) function. I formulate several frontier choice models and, using a pilot empirical study of transportation mode choice, compare the resulting estimates with those of the conventional logit specification. Most strikingly, it appears that the values of time implied by the frontier models are substantially below those of the logit model. This implies that policies designed to improve travel times may be of less value to consumers than is conventionally believed.     

The Location Choice of Employment‐based Immigrants among U.S. Metro Areas*

Abstract. This paper examines the initial location choice of legal employment‐based immigrants to the United States using Immigration and Naturalization Service data on individual immigrants, as well as economic, demographic, and social data to characterize the 298 metropolitan areas we define as the universal choice set. Focusing on interactions between place characteristics and immigrant characteristics, we provide multinomial logit model estimates for the location choices of about 38,000 employment‐based immigrants to the United States in 1995, focusing on the top 10 source countries. We find that, as groups, immigrants from nearly all countries are attracted to large cities with superior climates, and to cities with relatively well‐educated adults and high wages. We also find evidence that employment‐based immigrants tend to choose cities where there are relatively few immigrants of nationalities other than their own. However, when we introduce interaction terms to account for the sociodemographic characteristics of the individual immigrants, we find that the estimated effects of location destination factors can reverse as one takes account of the age, gender, marital status, and previous occupation of the immigrants.     

Evolving curricula and syllabi – challenges for physical geography

Recent technological changes in communications, in research and in learning and teaching styles should prompt developments in approaches to physical geography curricula. This paper questions whether there has been sufficient open discussion of such curricula and those that have been set nationally. Developments in England and Wales over more than fifty years provide the context for the present situation, which is considered according to types of knowledge under the headings of skills, concepts, and global and local knowledge. It is proposed that “Young Britannia”, a personification of the student who currently faces a changing world, deserves choices for physical geography training that provide sufficient alternatives. We suggest that there should be further open discussion of future curricular architectures for physical geography.     

Disaggregate Migration Behavior and the Volume of Interregional Migration

Nested multinomial logit models are used to investigate migration behavior during the 1971–74 period for a large sample of the population of Ecuador. The nested form of the model makes it possible to test hypotheses about the importance of destination characteristics in conditioning the odds for out-migration. Our empirical results indicate that the odds for migration from each origin are conditioned by the expected utilities of the available set of destinations, as well as characteristics of the origins and the personal characteristics of potential migrants. The association between destination characteristics and the frequency of out-migration allows the total volumes of migration to be adjusted to interregional differences in place-specific utilities.     

The Net Effects of Large Plant Locations and Expansions on County Employment*

This paper argues that the net economic impact of new firm locations or expansions is determined by a multitude of opposing forces. Using a unique database, I set out to evaluate the net effects of these opposing forces by looking at the net change in local employment and population arising from large (greater than 300 new jobs) firm locations or expansions in the State of Georgia. The analysis suggests that the employment multipliers associated with new firm locations are much less than one; that is, the net employment effect of a large firm opening is smaller than the gross employment impact. This result is consistent with other empirical economic impact studies, which find multipliers much smaller than those of typical input–output models, often less than unity, and a previous study showing little net effect of large plant openings. Expansions of existing establishments are shown to have substantial multiplicative effects, however, with an average employment multiplier of 2.0. I discuss possible reasons for differential impacts across new and expanding firms, focusing on the nature of the firms. Differences in net impact across industries and high‐tech versus low‐tech firms also is evaluated. I find that the impact of large firm locations or expansions on population in the resident county generally is negative, but positive for the broader region encompassing the county of location and its contiguous neighbors.     

The Spatial Distribution of Wages: Estimating the Helpman‐Hanson Model for Germany

Abstract.  Using German district data we estimate the structural parameters of a new economic geography model as developed by Helpman (1998) and Hanson (1998 , 2001a ). The advantage of the Helpman‐Hanson model is that it incorporates the fact that agglomeration of economic activity increases the prices of local (nontradable) services, like housing. This model thereby provides an intuitively appealing spreading force that allows for less extreme agglomeration patterns than predicted by the bulk of new economic geography models. Generalizing the Helpman‐Hanson model, we also analyze the implications for the spatial distribution of wages once the assumption of real wage equalization is dropped. If we no longer assume real wage equalization we find support for a spatial wage structure as well as for the relevance of the structural parameters of the theoretical model.