|
on Discrete Choice Models |
| By: | Michael P. Keane (University of Technology Sydney and Arizona State University); Robert M. Sauer (University of Southampton and IZA Bonn) |
| Abstract: | Two key issues in the literature on female labor supply are: (1) if persistence in employment status is due to unobserved heterogeneity or state dependence, and (2) if fertility is exogenous to labor supply. Until recently, the consensus was that unobserved heterogeneity is very important, and fertility is endogenous. But Hyslop (1999) challenged this. Using a dynamic panel probit model of female labor supply including heterogeneity and state dependence, he found that adding autoregressive errors led to a substantial diminution in the importance of heterogeneity. This, in turn, meant he could not reject that fertility is exogenous. Here, we extend Hyslop (1999) to allow classification error in employment status, using an estimation procedure developed by Keane and Wolpin (2001) and Keane and Sauer (2005). We find that a fairly small amount of classification error is enough to overturn Hyslop’s conclusions, leading to overwhelming rejection of the hypothesis of exogenous fertility. |
| Keywords: | female labor supply, fertility, discrete choice, classification error, simulated maximum likelihood |
| JEL: | J2 J6 C3 D1 |
| Date: | 2006–09 |
| URL: | http://d.repec.org/n?u=RePEc:iza:izadps:dp2332&r=dcm |
| By: | A. PRINZIE; D. VAN DEN POEL |
| Abstract: | Data mining applications addressing classification problems must master two key tasks: feature selection and model selection. This paper proposes a random feature selection procedure integrated within the multinomial logit (MNL) classifier to perform both tasks simultaneously. We assess the potential of the random feature selection procedure (exploiting randomness) as compared to an expert feature selection method (exploiting domain-knowledge) on a CRM cross-sell application. The results show great promise as the predictive accuracy of the integrated random feature selection in the MNL algorithm is substantially higher than that of the expert feature selection method. |
| Date: | 2006–05 |
| URL: | http://d.repec.org/n?u=RePEc:rug:rugwps:06/390&r=dcm |
| By: | Philip A. Haile (Department of Economics and Cowles Foundation, Yale University, and NBER); Ali Hortacsu (University of Chicago and NBER); Grigory Kosenok (NES) |
| Abstract: | The quantal response equilibrium (QRE) notion of McKelvey and Palfrey (1995) has recently attracted considerable attention, due in part to its widely documented ability to rationalize observed behavior in games played by experimental subjects. However, even with strong a priori restrictions on unobservables, QRE imposes no falsifiable restrictions: it can rationalize any distribution of behavior in any normal form game. After demonstrating this, we discuss several approaches to testing QRE under additional maintained assumptions. |
| Keywords: | quantal response equilibrium, falsifiability, testable restrictions, regular quantal response equilibrium, rank-cumulative probabilities, Block-Marschak polynomials |
| Date: | 2006–08 |
| URL: | http://d.repec.org/n?u=RePEc:cfr:cefirw:w0076&r=dcm |
| By: | Ivan Fernandez-Val (Department of Economics, Boston University); |
| Abstract: | Fixed e®ects estimates of structural parameters in nonlinear panel models can be severely biased due to the incidental parameters problem. In this paper I show that the most important com- ponent of this incidental parameters bias for probit ¯xed e®ects estimators of index coe±cients is proportional to the true parameter value, using a large-T expansion of the bias. This result allows me to derive a lower bound for this bias, and to show that ¯xed e®ects estimates of ratios of coe±cients and average marginal e®ects have zero bias in the absence of heterogeneity and have negligible bias relative to their true values for a wide range of distributions of regressors and individual e®ects. Numerical examples suggest that this small bias property also holds for logit and linear probability models, and for exogenous variables in dynamic binary choice models. An empirical analysis of female labor force participation using data from the PSID shows that whereas the signi¯cant biases in ¯xed e®ects estimates of model parameters do not contami- nate the estimates of marginal e®ects in static models, estimates of both index coe±cients and marginal e®ects can be severely biased in dynamic models. Improved bias corrected estimators for index coe±cients and marginal e®ects are also proposed for both static and dynamic models. |
| JEL: | C23 C25 J22 |
| Date: | 2005–10 |
| URL: | http://d.repec.org/n?u=RePEc:bos:wpaper:wp2005-38&r=dcm |
| By: | Brodaty, Thomas; Gary-Bobo, Robert J.; Prieto, Ana |
| Abstract: | We propose to model individual educational investments as a rational decision, maximizing expected utility, conditional on some characteristics observed by the student, under the combined risks affecting future wages and schooling duration. Assuming that students' attitudes toward risk can be represented by a CRRA utility, we show that the risk-aversion parameter can be identified in a natural way, using the variation in school-leaving ages, conditional on certified educational levels. Estimation can be performed by means of classic Maximum Likelihood methods. The model can easily be compared with a non-structural, simplified version, which is a standard wage equation with endogenous dummy variables representing education levels, education levels being themselves determined by an Ordered Probit model. We find small but significant values of the coefficient of relative risk aversion, between 0.1 and 0.9. These results are obtained with a rich sample of 12,500 young men who left the educational system in 1992, in France. |
| Keywords: | econometrics; human capital; returns to education; risk aversion |
| JEL: | I2 J24 J31 |
| Date: | 2006–06 |
| URL: | http://d.repec.org/n?u=RePEc:cpr:ceprdp:5694&r=dcm |
| By: | Drago, Francesco; Kadar, Dora |
| Abstract: | In this paper we develop a theory of time-inconsistency and regret that is motivated by evidence on a 'price discrimination' technique widespread in the United States, namely mail-in-rebate promotions. Our model combines partial naivete about future self-control problems and the sunk-cost effect (regret). We assume that agents deviating from their past choices suffer a certain emotional disutility from having brought a bad decision in the past and that this emotional disutility is negatively related to the length of the period between the choice made and the deviation from it. In the context of our application the model explains why in a multi period setting a large number of consumers respond to the rebate offers intending to redeem the rebate and then fail to provide the necessary effort when it comes to collect their money. Moreover, consumer failure to accomplish a task planned in the past (e.g. redeeming the rebate) is more likely when the deadline of completion is longer. This prediction is supported by experimental studies on various forms of procrastination and by field and experimental evidence on mail-in-rebates. We review a number of areas for which the theory may have important implications. |
| Keywords: | mail-in-rebate; naivete; regret; time inconsistency |
| JEL: | C70 D11 D91 |
| Date: | 2006–07 |
| URL: | http://d.repec.org/n?u=RePEc:cpr:ceprdp:5757&r=dcm |
| By: | Jawwad Noor (Department of Economics, Boston University) |
| Abstract: | Due to factors such as temptation, choices may not respect normative prefer- ence (the agent.s own, subjective view of what constitutes his welfare). Neverthe- less, the evidence on preference reversals suggests a means of recovering normative preference from choice. A de.nition of normative preference in terms of choices between su¢ ciently delayed alternatives is formulated and studied. Mild condi- tions on behavior are shown to ensure the existence of a normative preference. Two characterizations are provided. It is demonstrated that a notion of welfare may exist inspite of dynamic inconsistency of preferences. An application shows that the evidence on hyperbolic discounting implies that agents.normative discount functions must be exponential. |
| Keywords: | Preference Reversals, Welfare, Temptation, Hyperbolic Discount- ing, Ethics. |
| JEL: | D11 |
| Date: | 2005–10 |
| URL: | http://d.repec.org/n?u=RePEc:bos:wpaper:wp2005-039&r=dcm |
| By: | Jawwad Noor (Department of Economics, Boston University); |
| Abstract: | Choice may be determined both by a consideration of one?s welfare (normative preference) and by desires (temptation preference). To provide foundations for such a theory, Gul and Pesendorfer [7, 8] adopt a preference over choice problems as a primitive and hypothesize that temptation creates a preference for commit- ment. This paper argues that temptation may in fact create the absence of a preference for commitment, and that the primitive may not be empirically mean- ingful since it requires us to observe behavior in the absence of temptation. An alternative approach to providing foundations is introduced. Motivated by the evidence on preference reversals, it is hypothesized that delayed temptations are easier to resist than immediate temptations. Normative preference is derived via choices between su¢ ciently delayed alternatives, and temptation preference is in- ferred from discrepancies between normative preference and choice. With a choice correspondence as the primitive, agents who are ?tempted not to commit?are mod- eled. The foundations of the model are used to identify evidence supporting such temptation. |
| Keywords: | Self-Control, Temptation, Commitment, Preference Reversals, Re- vealed Preference. |
| JEL: | D11 D60 |
| Date: | 2006–04 |
| URL: | http://d.repec.org/n?u=RePEc:bos:wpaper:wp2006-019&r=dcm |
| By: | Marie-Louise Vierø (Department of Economics, Queen's University) |
| Abstract: | This paper derives a representation of preferences for a choice theory with vague environments; vague in the sense that the agent does not know the precise probability distributions over outcomes conditional on states. Instead, he knows only a possible set of these probabilities for each state. Thus, the paper relaxes an assumption about the environment, which is done while maintaining the independence axiom. The behavior implied by this model is different from both the behavior implied by standard subjective expected utility models and the behavior implied by ambiguity aversion models. To illustrate these differences and the importance of allowing for vagueness, the consequences of the developed theory for a simple contracting problem are considered. The paper also provides a defense of the independence axiom against the usual Ellsberg critique. |
| Keywords: | Decision Theory, Vagueness, Utility, Optimism, Positivism |
| JEL: | D80 D81 D00 |
| Date: | 2006–10 |
| URL: | http://d.repec.org/n?u=RePEc:qed:wpaper:1094&r=dcm |