nep-dcm New Economics Papers
on Discrete Choice Models
Issue of 2005‒05‒14
eleven papers chosen by
Philip Yu
Hong Kong University

  1. Análisis del nuevo acuerdo de capitales de Basilea (BIS-II): PYME-risk, country-risk y operational-risk By Mariano González
  2. Some Practical Guidance for the Implementation of Propensity Score Matching By Marco Caliendo; Sabine Kopeinig
  3. Non-Market Values and Intra-Household Gender Gap in Healthcare: The Case of Rural China By Mengtao Gao; Yang Yao
  4. A Stronger Case for Transitive Preferences By Theodore Turocy
  5. O Trabalho Infantil na Ukrania By Manuela Magalhaes
  6. Modeling the Choice of Telecommuting 2: A Case of the Preferred Impossible Alternative By Patricia Mokhtarian; Ilan Salomon
  7. Modeling the Choice of Telecommuting 3: Identifying the Choice Set and Estimating Binary Choice Models for Technology-Based Alternatives By Patricia Mokhtarian; Ilan Salomon
  8. The precautionary principle in environmental policy and the theory of choice under uncertainty By John Quiggin
  9. Increasing Uncertainty: A Definition By Simon Grant; John Quiggin
  10. Exchange Rates and FOMC Days By Michael Melvin
  11. Designer Path Independent Choice Functions By Mark Johnson; Richard Dean

  1. By: Mariano González
    Abstract: trabajo analiza el Nuevo Acuerdo de Capitales del Banco Internacional de Pagos de Basilea, conocido como BIS-II, en lo relativo al riesgo de crédito con Pymes, riesgo país y riesgo operacional. Estudia los conceptos, clasificaciones y métodos de estimación propuesto por la norma, así como las ventajas e inconvenientes de cada una de las opciones. Presenta además, para cada uno de los tres riesgos estudiados, una propuesta de rating basada en el estudio de muestras diferentes para cada caso, Pymes españolas, información sobre países suministrada por el Banco Mundial y entidades financieras españolas. Finalmente, recoge un modelo econométrico que cumple las exigencias de BIS-II para modelos avanzados en los respectivos riesgos estudiados, y que permite incluir tanto efectos aleatorios como correlación, espacial y temporal. Todo ello se acompaña con una abundante y actualizada bibliografía sobre los temas tratados.
    Keywords: Basel-II, Credit risk, Country risk, Operational risk, rating, Dynamic Ordered probit, Simulated Maximum Likelihood.
    JEL: C15 C23 C25 C61 E22 F34 G21 G33 H
    Date: 2004–07–29
  2. By: Marco Caliendo (DIW Berlin and IZA Bonn); Sabine Kopeinig (University of Cologne)
    Abstract: Propensity Score Matching (PSM) has become a popular approach to estimate causal treatment effects. It is widely applied when evaluating labour market policies, but empirical examples can be found in very diverse fields of study. Once the researcher has decided to use PSM, he is confronted with a lot of questions regarding its implementation. To begin with, a first decision has to be made concerning the estimation of the propensity score. Following that one has to decide which matching algorithm to choose and determine the region of common support. Subsequently, the matching quality has to be assessed and treatment effects and their standard errors have to be estimated. Furthermore, questions like "what to do if there is choice-based sampling?" or "when to measure effects?" can be important in empirical studies. Finally, one might also want to test the sensitivity of estimated treatment effects with respect to unobserved heterogeneity or failure of the common support condition. Each implementation step involves a lot of decisions and different approaches can be thought of. The aim of this paper is to discuss these implementation issues and give some guidance to researchers who want to use PSM for evaluation purposes.
    Keywords: propensity score matching, implementation, evaluation, sensitivity
    JEL: C40 H43
    Date: 2005–05
  3. By: Mengtao Gao (CCER - China Center for Economic Research); Yang Yao (CCER - China Center for Economic Research)
    Abstract: This paper studies the age structure of the gender gap in household health care allocation by using survey data coillected on 1428 rural households (8414 persons) in 8 Chinese provinces. The primary concerns are rthe treatment rate and expenditure conditional on reported 2-week illnesses. To avoid the potential bias in self-reported illness, in particular, the bias arising from wome's tendency to report more illnesses than men, conditional probit and OLS analyses are adopted. in addition, to take care the possibility that some illnesses have different impacts on men and women, we suppplement the study by looking at people's responses to two specific illnesses, cold and diarrhoea, that do not have gender implications. We employ several sets of variables measuring a person's work capability, occupation, political affiliation, and education to control his potential market values. Our results show that girls under age 13 do get significantly less medical care than boys of the same age, but prime-age wives get more than the husbands, and old-age wives get less than old-age husbands. While the results for the children agree with other studies, the pattern for prime-age and old-age adults is new and consistent with the considerations proposed above.
    Keywords: China, health care allocation, gender, rural
    JEL: I31 I38
    Date: 2004–10
  4. By: Theodore Turocy (Texas A&M University)
    Abstract: The assumption that preferences are transitive, or, equivalently, that choice behavior satisfies the Weak Axiom of Revealed Preference, is at the core of most economic theory. While this is a natural assumption, one could ask the degree to which it is restrictive: are there objectives that could not be attained by such behavior that could be attained by choices violating the assumption? It is argued that the answer to this question is no in one setting of choice under random budget sets.
    Keywords: transitivity
    JEL: C7 D8
    Date: 2005–05–10
  5. By: Manuela Magalhaes
    Abstract: As crianças fazem parte da vida económica das sociedades. A evidência histórica mostra que o contributo das crianças para o rendimento das famílias é consistentemente significativo. O trabalho infantil é, ainda nos dias de hoje, um fenómeno expressivo, quer nas sociedades menos desenvolvidas, quer nas mais desenvolvidas. Este estudo analisa os factores determinantes na decisão das famílias, relativamente á afectação do tempo das crianças entre a escola; a escola e o trabalho, o trabalho ou o lazer. O modelo logit multinomial é o modelo utilizado para identificar as características pessoais, familiares, locais e sazonais que influenciam a decisão das famílias Ukranianas. Os resultados mostram que todos estes factores são considerados pelas famílias nas suas decisões. Assim, políticas que promovam o desenvolvimento económico, a redução do agregado familiar e alternativas à escola no terceiro trimestre, podem contribuir significativamente para a redução do trabalho infantil na Ukrania
    Keywords: Trabalho Infantil, Logit Multinomial
    JEL: J
    Date: 2005–05–12
  6. By: Patricia Mokhtarian (University of California, Davis); Ilan Salomon (Hebrew University)
    Abstract: A conceptual model of the choice to telecommute was advanced in an earlier paper (Mokhtarian and Salomon, 1994). In this paper, we present empirical data from a non-representative sample of 628 City of San Diego employees on key variables and relationships in that model. The relationships among possibility, preference, and choice are examined. A key finding is the existence of a large group of people (57% of the sample) for whom telecommuting is a Preferred Impossible Alternative. Dichotomous and continuous constraints are distinguished, and three dichotomous constraints are defined. Lack of awareness is active for 4%, job unsuitability for 44%, and manager disapproval for 51% of the sample. For 68% of the sample, at least one of these constraints is active. Even among those for whom none of the dichotomous constraints is in force, most people do not choose telecommuting due to the presence of active continuous constraints. For only 11% of the entire sample, telecommuting is possible, preferred, and chosen. The potential impacts of self-selection bias are estimated, and sampling bias is qualitatively assessed. This analysis provides a crude but useful estimate of the potential of telecommuting in the population, and more specifically, the relative share of potential telecommuters who are prevented by key dichotomous constraints from choosing that option.
    Keywords: telecommuting, teleworking
    JEL: J
    Date: 2005–05–12
  7. By: Patricia Mokhtarian (University of California, Davis); Ilan Salomon (University of California, Davis)
    Abstract: Previous papers in this series have presented a conceptual model of the individual decision to telecommute and explored relationships among constraints, preference, and choice. A related paper has developed a binary model of the preference for home-based telecommuting. Noting that there is a wide gap between preferring to telecommute (88% of the sample) and actually telecommuting (13%), this paper develops binary logit models of telecommuting adoption. Two approaches to dealing with constraints are compared: incorporating them directly into the utility function, and using them to define the choice set. Models using the first approach appear to be statistically superior in this analysis, explaining 63-64% of the information in the data. Variables significant to choice include those relating to work and travel drives, and awareness, manager support, job suitability, technology, and discipline constraints. The best model was used to analyze the impact of relaxing three key constraints on the 355 people in the sample for whom telecommuting was previously identified to be a Preferred Impossible Alternative. When unawareness, lack of manager support, and job unsuitability constraints are relaxed, 28% of the people in the PIA category would be expected to adopt telecommuting. The importance of behavioral models to accurately forecasting telecommuting adoption is emphasized and is suggested to have wider implications for predicting technology-based activity changes.
    Keywords: telecommuting, teleworking, discrete choice, choice set
    JEL: J
    Date: 2005–05–12
  8. By: John Quiggin (Risk & Sustainable Management Group, School of Economics, University of Queensland)
    Abstract: The precautionary principle, presented as a guide to environmental policy decisions in the presence of uncertainty, has been the subject of vigorous debate. However, the has generally not been discussed in relation to formal theories of choice under uncertainty developed as generalizations of the expected utility model. In this paper, it is argued that a formal basis for the precautionary principle may be found in an incompleteness hypothesis regarding formal models of choice under uncertainty. The incompleteness hypothesis states that estimates derived from formal models of choice under uncertainty will generally be over-optimistic and that the errors will be greater, the less well-understood is the problem in question.
    Keywords: precautionary principle, generalized expected utility theory
    JEL: D81 Q2
    Date: 2005–04
  9. By: Simon Grant (Department of Economics, Rice University); John Quiggin (Department of Economics, University of Queensland)
    Abstract: We present a definition of increasing uncertainty, in which an elementary increase in the uncertainty of any act corresponds to the addition of an `elementary bet' that increases consumption by a fixed amount in (relatively) `good' states and decreases consumption by a fixed (and possibly different) amount in (relatively) `bad' states. This definition naturally gives rise to a dual definition of comparative aversion to uncertainty. We characterize this definition for a popular class of generalized models of choice under uncertainty.
    Keywords: uncertainty, ambiguity, risk, non-expected utility
    JEL: C72 D81
    Date: 2004–05
  10. By: Michael Melvin (W. P. Carey School of Business Department of Economics)
    Abstract: FOMC meeting days provide a natural laboratory for exploring the effects of policy uncertainty and learning on exchange rate determination. Intradaily mark/dollar exchange rates are employed for 10 FOMC meetings. The meetings examined are the first 10 following the February 1994 change in policy where the meeting outcome is announced after meetings end. The following hypotheses motivated by the market microstructure literature are examined: 1) strategic behavior by informed traders should result in position-taking prior to meeting end and the revelation of policy and 2) bid-ask spreads should widen due to adverse selection potential as the probability of quoting to an informed trader increases. A markov-switching model is used to estimate the time of informed position-taking. The data suggest that on most days, there is a switch to the informed-trading state during the time of the meeting, well before the end of the meeting. An extensive search of public news indicates that the informed trading cannot be explained as the response to public information. An ordered probit model of the bid-ask spread is estimated as a function of the probability of being in the informed trading state. The estimation results indicate that the greater the probability of being in the informed trading state, the wider spreads. This is consistent with dealers protecting against adverse selection in quoting. The evidence indicates that meeting outcomes are generally anticipated during the meeting. In this sense, the realization of the meeting outcome is often not news.
  11. By: Mark Johnson (W. P. Carey School of Business Department of Economics); Richard Dean (California Institute of Technology - Department of Mathematics)
    Abstract: This paper provides a new characterization result for path independent choice functions (PICF) on finite domains and uses that characterization as the basis of an algorithm for the construction of all PICFs on a finite set of alternatives, V, designed by an a priori given set I of initial choices as well as the determination of whether the initial set I is consistent with path independence. The characterization result identifies two properties of a partition of the Boolean algebra as necessary and sufficient for a choice function C to be a PICF: (i): For every subset A of V the set arc(A) = {B: C (B) = C(A)} is an interval in the Boolean algebra 2v. (ii): If A/B is an interval in the Boolean algebra such that C(A) = C(B) and if M/N is an upper transpose of A/B then C(M) = C(N). The algorithm proceeds by expanding on the implications of these two properties.

This nep-dcm issue is ©2005 by Philip Yu. It is provided as is without any express or implied warranty. It may be freely redistributed in whole or in part for any purpose. If distributed in part, please include this notice.
General information on the NEP project can be found at For comments please write to the director of NEP, Marco Novarese at <>. Put “NEP” in the subject, otherwise your mail may be rejected.
NEP’s infrastructure is sponsored by the School of Economics and Finance of Massey University in New Zealand.