This research explores the medical care consumption and absenteeism decisions of employed individuals with acute illnesses in an effort to better understand health care behavior. Using data from the 1987 National Medical Expenditure Survey, the author estimates the structural parameters of an individual's discrete choice stochastic optimization problem, as opposed to employing conventional reduced form estimation methods that are prevalent in the health care literature. The estimates allow for predictions of the change in physician services use and illness-related absenteeism that arise with the introduction of new public policy initiatives involving health insurance and sick-leave coverage.
Many non/semiparametric time series estimates may be regarded as different forms of sieve extremum estimates. For stationary absolute regular mixing observations, the authors obtain convergence rates of sieve extremurn estimates and root-n asymptotic normality of 'plug-in' sieve extremum estimates of smooth functionals. As applications to time series models, they give convergence rates for nonparametric ARX(p, q) regression via neural networks, splines, wavelets; root-n asymptotic normality for partial linear additive AR(p) models, and monotone transformation AR(1) models.
This paper proposes a new statistical model for the analysis of data which arrives at irregular intervals. The model treats the time between events as a stochastic process and proposes a new class of point processes with dependent arrival rates. The conditional intensity is developed and compared with other self-exciting processes. The model is applied to the arrival times of financial transactions and therefore is a model of transaction volume, and also to the arrival of other events such as price changes. Models for the volatility of prices are estimated, and examined from a market microstructure point of view.
This study uses Social Security data on the earnings of military applicants to the all-volunteer forces to compare the earnings of Armed Forces veterans with the earnings of military applicants who did not enlist. Matching, regression, and Instrumental Variables (IV) estimates are presented. The matching and regression estimates control for most of the characteristics used by the military to select qualified applicants from the military applicant pool. The IV estimates exploit an error in the scoring of exams used by the military to screen applicants between 1976 and 1980. All the estimates suggest that soldiers who served in the early 1980s were paid considerably more than comparable civilians while in the military. Military service also appears to have led to a modest (less than 10 percent) increase in the civilian earnings of nonwhite veterans while actually reducing the civilian earnings of white veterans. Most of the positive effects of military service on civilian earnings appear to be attributable to improved employment prospects for veterans.
In Aumann [1987], it is asserted that for those who adhere to the “...Bayesian view of the world, the notion of equilibrium is an un-avoidable consequence... ” I discuss two possible interpretations of the information model and show that neither interpretation supports this assertion. The hierarchy representation interpretation renders the prior stage meaningless and hence both the key assumption of Au-mann’s theory and its conclusion become impossible to interpret. The prior interpretation, on the other hand, is distinctively non-Bayesian. Furthermore, both the common prior assumption and the notion of having beliefs over one’s own actions are problematic in the latter interpretation. Key Words: correlated equilibrium, Bayesian view of probability, common prior assumption, Savage’s foundations of statistics. † I am indebted to Eddie Dekel-Tabak, David Kreps, and Robert Wilson for their help. Financial sup-port from the Alfred P. Sloan Foundation and the National Science Foundation is gratefully acknowledged. Aumann [1987] presents a result which he interprets as establishing that, for those who adhere to the “...Bayesian view of the world, the notion of equilibrium is an unavoidable consequence... ”1 The purpose of this comment is to refute this claim. An information model I = Ω, (Ti, pi)
IMPARTIALITY IS THE MORAL IMPERATIVE requiring that conflicting claims be evaluated without prejudice. In this paper I propose an axiomatic definition of impartiality and examine its implications for the theory of social welfare functions. Following the seminal work of Harsanyi (1953, 1977) I take as given the set of individuals that constitute the society and the set of social alternatives, representing the constitutions, income distributions, institutions, or policies among which the society must choose. Moreover, the moral value judgment that should govern this choice is modeled as a preference relation of an ethical observer. However, unlike Harsanyi, who defines the observer's preference relation on the set of all extended lotteries (i.e., joint probability distributions on social alternatives and individuals) I define the observer's preference relation on a set of allocations whose elements are assignments of social-alternative lotteries (i.e., probability distributions on the set of social alternatives) to individuals. As in Harsanyi's theory, the observer's preference relation is supposed to govern the choice among social alternatives of individuals placed behind a veil of ignorance regarding their social position and preferences. Harsanyi assumes that the observer's preference relation on the set of extended lotteries and the individual preference relations on the set of social-alternative lotteries satisfy the axioms of expected utility theory and jointly satisfy the principle of acceptance.2 He shows that the observer's preference relation may be represented as a weighted sum of individual von Neumann-Morgenstern utilities and defines impartiality as the restriction that the individual utilities are assigned equal weights. This representation may be interpreted as assigning equal probabilities to the events of being each individual in society. Note, however, that given any preference relation that is representable as a weighted sum of individual utilities with strictly positive weights, a new set of individual utilities may be defined (by multiplying each utility function by its weight and dividing through by the inverse of the number of individuals) to obtain a new representation of the preference relation with uniform weights. In other words, the same observer's preference relation is represented as a weighted sum of individual utilities with equal weights. Since this manipulation does not change the underlying observer's preference relation, it does not make it impartial except in a tautological sense.
The least-absolute-deviations (LAD) estimator for a median-regression model does not satisfy the standard conditions for obtaining asymptotic refinements through use of the bootstrap because the LAD objective function is not smooth. This paper overcomes this problem by smoothing the objective function. The smoothed estimator is asymptotically equivalent to the standard LAD estimator. With bootstrap critical values, the rejection probabilities of symmetrical t and X 2 tests based on the smoothed estimator are correct through O(n -γ ) under the null hypothesis, where γ<1 but can be arbitrarily close to 1. In contrast, first-order asymptotic approximations make errors of size O(n -γ ). These results also hold for symmetrical t and X 2 tests for censored median regression models.
This paper attemps to identify, in a framework deliberately stripped of unnecessary technical- ities, some of the basic reasons why adaptive learning may or may not lead to stability and convergence to self-fulfilling expectations in large socioeconomic systems where no agent, or collection of agents, can act to manipulate macroeconomic outcomes. It is shown that if agents are somewhat uncertain about the local stability of the system, and are accordingly ready to extrapolate a large range of regularities (trends) that may show up in past small deviations from equilibrium, including divergent ones, the learning dynamics is locally divergent. On the other hand, if agents are fairly sure of the local stability of the system, and extrapolate only convergent trends out of small past deviations from equilibrium, one may get local stability. This “uncertainty principle” does show up in a wide variety of contexts: smooth or discontin- uous, finite or infinite memory learning rules, error learning, recursive least squares, Bayesian learning.
This paper concerns inferring how self-interested subjects, as opposed to altruistic investigators, evaluate treatments in social experiments. The authors argue that the attrition behavior of subjects reveals their evaluation and discuss the usefulness of using such data in performing subject-based evaluation. The authors study the causes of disagreements between investigators and subjects in evaluating treatments and empirically assess the degree to which they disagree. The paper provides an empirical framework for estimating the systematic level of disagreement in the presence of such errors. Using clinical trials, the authors find substantial evidence of overapproval by investigators in about one-third of the trials analyzed.
The method of simulated scores (MSS) is presented for estimating limited dependent variables models (LDV) with flexible correlation structure in the unobservables. We propose simulators that are continuous in the unknown parameter vectors, and hence standard optimization methods can be used to compute the MSS estimators that employ these simulators. The first continuous method relies on a recursive conditioning of the multivariate normal density through a Cholesky triangularization of its variance-covariance matrix. The second method combines results about the conditionals of the multivariate normal distribution with Gibbs resampling techniques. We establish consistency and asymptotic normality of the MSS estimators and derive suitable rates at which the number of simulations must rise if biased simulators are used.