Hashem Dezhbakhsh

On the Presence of Speculative Bubbles in Stock Prices

We examine empirically the existence of speculative bubbles in U.S. stock prices and, by building on Wests procedure, propose direct and computationally simple tests of the “nobubble†hypothesis. These tests are likely to be close to their nominal size in small samples and to have small sample power against a wide class of bubbles including those orthogonal to the dividend process. We apply the tests to long-term annual U.S. stock market data for the 1871–1981 and 1871–1988 periods. Contrary to West, we do not reject the “no-bubble†hypothesis. We offer an explanation as to the cause of discrepancy between the two results.

The Inappropriate use of Serial Correlation Tests in Dynamic Linear Models

A survey of several economic journals reveals that very often the Durbin-Watson and the portmanteau (Box-Pierce or Ljung-Box) tests are inappropriately applied to linear models with lagged dependent variables and exogenous regressors. Sampling experiments indicate that the Durbin-Watson performs poorly in models with more than one lag of the dependent variable, a situation commonly considered in the literature. The experiments also indicate that the portmanteau test is inadequate when applied to dynamic linear models with exogenous regressors. In addition, the performance of Durbins h and m tests in models commonly used in the literature but not considered by previous studies is evaluated. The results reveal that among the four tests examined, the one which is the least frequently used in practice (the m test) has the best performance. Copyright 1990 by MIT Press.

The effect of concealed handgun laws on crime: beyond the dummy variables

So far 33 states have adopted right-to-carry concealed handgun laws. The advocates argue these laws have a deterrent effect on crime, while the opponents believe they facilitate crime by increasing gun availability. Although both sides assume that these laws affect behavior, no attempt has yet been made to model such effects using crime theory. Consequently, the empirical evidence on such effects lack a theoretical basis; for example, a highly publicized study by Lott and Mustard (1997) inappropriately models the effect of the law through a dummy variable (a binary-valued regressor). We extend the economic model of crime to formulate a theoretical basis for empirical examination of the issue. We show that using a dummy variable leads to misspecification, and use an alternative procedure to estimate the effect of concealed handgun laws in 1992 for states which had not yet adopted such laws. Our results show that the expected effect of the law on crime varies across the counties and states and depends on county-specific characteristics in a meaningful way. Such effects appear to be much smaller and more mixed than Lott and Mustard suggest, and are not crime-reducing in most cases.

On the typical spectral shape of an economic variable

In a classic article, Granger (Econometrica 34, 1966) asserted that most economic time series measured in level have spectra that exhibit a smooth declining shape with considerable power at very low frequencies. There has been no systematic attempt to examine Grangers assertion with international data. Output level spectra are estimated for 58 countries, divided into developed, high-income developing, and low-income developing groups. The shapes of the estimated spectra are found to be strikingly similar to Grangers typical shape, particularly for the developed countries.

Foreign Exchange Forward and Futures Prices: Are They Equal?

Small sample t -test results are reported in the literature that indicate the difference between futures and forward exchange rates are statistically insignificant. Much research draws on this finding, which is in contrast with theory. The evidence presented here suggests this difference does not follow a normal distribution, so the small sample inferences based on the t -tests may be suspect. As appropriate alternatives, nonparametric distribution-free tests are used to reexamine the difference for two sample periods, one covering the 1970s and the other the 1970s and 1980s. A significant divergence is observed for several currencies as well as for a sample of pooled currencies. The results are stronger for samples covering the 1980s. The economic significance of the forward-futures differentials is examined and theoretical justifications are discussed.

Testing for autocorrelation in the presence of lagged dependent variables: A specification error approach

Abstract Hausmans specification error testing procedure is used to develop serial correlation tests in lagged dependent variable models. Properties of the tests are discussed and comparisons are made with existing tests. A reparameterization of the model using the error structure under the alternative hypothesis results in an augmented regression equation; standard tests of exclusionary restrictions on coefficients of the augmented regressors can be regarded as serial correlation tests. Monte Carlo experiments are used to examine small sample properties. Several of the newtests are shown to perform well in comparison with existing tests.

Enrollment at Highly Selective Private Colleges: Who is Left Behind?

This study examines the enrollment decisions of freshmen applicants to highly selective private colleges relative to highly selective public colleges. The empirical analysis shows that college enrollment varies by family income - in particular, low-income students with demonstrated financial need are less likely to enroll at private colleges, controlling for other factors including race and financial aid. Grant aid tends to have a moderating effect but the amount of aid under the existing college pricing and financial aid system appears to be inadequate. This finding provides support for recent revamping of financial aid offerings to low-income students by some highly selective private colleges. In addition, the results suggest that while family income appears to be a strong determinant of enrollment at these colleges, race does not seem to play a role.

Equivalence properties of the Hausman statistic based on a generalized inverse

Abstract The Hausman statistic based on a generalized inverse is numerically equivalent to statistics that are based on a subset of the parameters and do not require the use of a generalized inverse. In a general non-linear setting there exists at least one such statistic. In the case of the linear model a number of such statistics are obtained by arbitrarily selecting parameter subsets.

Strategic competition in the banking industry

A model of interbank rivalry is developed and tested in the context of a one period duopolistic game with stochastic variability in deposit and loan demand. A procedure involving two specification tests is used to examine model implications. Results confirm that banks change their lending rates not only in response to changing own cost conditions but also in recognition of mutual interdependence – strategic competition. The possibility that banks respond strategically to each other may have implications for antitrust analysis.

Are selective private and public colleges affordable

We examine college affordability under the existing pricing and financial aid system that awards both non need-based and need-based aid. Using data of freshmen attending a large number of selective private and public colleges in the USA, we find that the prices students actually pay for college have increased over time. Need-based grant aid has not kept pace with the substantial increases in non need-based aid. Most importantly, although low-income students received more subsidies than higher-income students, the existing financial aid system does not provide enough affordability to needy students. Nonetheless, the deficiency cannot be attributed to the increases in non need-based aid.