Search

This paper presents the importance of the visual expression of factor decomposition in regression analysis, which is particularly worthwhile for undergraduate students whose majors are not mathematics but social science. The conventional purpose of regression analysis is to examine specific hypotheses empirically. In particular, the statistical significance of the explanatory variable was tested, which may have been difficult for many students to understand mathematically. To remedy this, factor decomposition is introduced in the same way that human body composition is broken down into water, fat, and muscle. As an illustrative example, multiple regression was applied to the determinants of housing rents in Japan. The explanatory variables were the living area, building age, and walking time from the nearest station. The findings suggest that, with the help of visual expression, a student can easily appreciate which variable significantly affects housing rents.

In statistics classes, the central limit theorem has been demonstrated using simulation‐based illustrations. Known population distributions such as a uniform or exponential distribution are often used to consider the behavior of the sample mean in simulated samples. Unlike such simulations, a number of real‐data‐based simulations are here implemented in which the populations are empirical distributions of data selected from Japanese firms. The dataset chosen contains 38 variables familiar to business students, such as sales and assets. The maximum population size of the variables is 2243. One thousand samples with replacement are selected for specific variable–sample size combinations. Hypothesis testing results indicate that the normality hypothesis for the sample mean is rejected for 31 variables at the 0.1% level even with a sample size of 500. It is emphasized that the data for these variables indicate that this should not be a surprise, and emphasize the importance of looking at data. [ABSTRACT FROM AUTHOR]

Corporate finance data, such as cash holdings and dividend payments, are simultaneously determined. In this study, new endogeneity was introduced into a simultaneous equation model. The proposed model features a correlation between the explanatory variable in the first equation and the disturbance term in the second equation. The necessity of this endogeneity can be derived from a theoretical requirement and/or model identification requirement. Covariance structure analysis was performed to estimate this model using the maximum likelihood method. The efficacy of the proposed model is assessed by comparing it with conventional estimation methods, such as two-stage least squares. Finally, a Monte Carlo simulation was conducted to determine the bias conventional estimation methods produce in the case of this new endogeneity. [ABSTRACT FROM AUTHOR]

The article explains the two related-variables selection methods for temporal disaggregation. The effectiveness of these methods is tested by way of simulations and empirical applications.

The special and limiting cases of skewed generalized t distribution developed by Theodossiou (1998) include 11 alternative distributions. This study proposes likelihood ratio tests to select the best distribution from among 12 alternatives. The efficacy of this approach is shown through empirical applications and two kinds of Monte Carlo simulations. Daily stock returns data from 1,045 Japanese firms offer the following evidence. Student's t distribution is selected for 42% of the firms, generalized t distribution is selected for 35% of the firms, and skewed generalized t distribution is selected for 14% of the firms. On the one hand, non-data-based simulation provides evidence that the proposed procedure performs well when individual distributions are clearly separated. On the other hand, data-based simulation shows that a very large sample (e.g., 10,000) is needed to ensure good performance of the proposed procedure. [ABSTRACT FROM AUTHOR]

To purchase or authenticate to the full-text of this article, please visit this link: http://dx.doi.org/10.1111/j.0307-3378.2008.00276.x Byline: Kosei Fukuda (*) Keywords: age-earnings profile; age-period-cohort decomposition; Bayesian cohort model Abstract: ABSTRACT Aggregate data on US earnings, classified by period and by age, are decomposed into age, period and cohort effects, using the Bayesian cohort models, which were developed to overcome the identification problem in cohort analysis. The main findings, obtained by comparing college and high school graduates, are threefold. First, the age effects show a downward trend for the age group of 45-49 onwards for high school graduates but do not show any such trend for college graduates. Second, the period effects show a downward trend for high school graduates but reveal no such trend for college graduates. Third, the cohort effects are negligible for both college and high school graduates. Author Affiliation: (*)College of Economics, Nihon University, Japan Article note: Correspondence: College of Economics, Nihon University, 1-3-2, Misakicho, Chiyoda-ku, Tokyo 101-8360, Japan. Tel: +81 3 3219 3458; Fax: +81 3 3219 3458; Email: fukuda@eco.nihon-u.ac.jp. I am grateful to two anonymous referees for their very helpful comments and suggestions. In particular, the extension of Thornton et al.'s (1997) data sets would have been impossible without the referee's suggestion. Needless to say, any remaining errors are mine.

A forecasting model for real-time performance that allows for data revisions is presented. A modeling approach was used, and an observed time series was decomposed into stochastic trend, observation noise, and data revision in real time. It is concluded that this approach answered the need for data revision, which was not present in the conventional model developed by AC Harvey and EP Howrey.

A nation's economy switches from one regime to another; this feature is known as the asymmetry of business cycles. The Markov switching model is applied to forecast economic time series. These regime switches may arise between stationary and nonstationary processes. Stochastic unit root models are potentially useful in multi-step-ahead forecasting.

It is widely recognized that financial data distributions have heavier tails than normal distributions. These tails can be modeled as Student distributions and normal mixture distributions. However, selecting between these distributions has not been previously studied. To this end, this study proposes an information-criterion approach. The efficacy of the proposed method for selecting the number of mixture components is examined via simulation studies that compare the conventional hypothesis test. Empirical results using Japanese stock returns data are also provided. [ABSTRACT FROM AUTHOR]

In this study, a world diffusion index is developed to measure how uncertainty shocks have diffused among 179 economies and caused contractions in the world growth cycles. This index is simply defined as the percentage share of the number of expanding countries. It identifies four uncertainty shocks: the oil crisis of 1973; the bursting of the information technology bubble in 2000; the credit crunch of 2007; and the European debt crisis of 2010. To overcome the problem of data unavailability in emerging market economies, the annual GDP values of 179 economies are transformed through temporal disaggregation, and the dating of quarterly growth cycles is implemented as per the OECD method. The empirical findings indicate that each of the uncertainty shocks caused severe contractions in the advanced economies but that the emerging market economies experienced such contractions only during the credit crunch of 2007. Policy implications are also discussed. [ABSTRACT FROM AUTHOR]

This study proposes a model selection approach for determining the inclusion or exclusion of a latent variable when two exogenous and two endogenous variables are provided. The models compared are the multivariate regression model without latent variables (MR model) and the multiple indicators multiple causes model (MIMIC model). The inclusion of a latent variable in the MR model yields the MIMIC model. In the proposed approach, an information criterion is used to select the best model of the two. The efficacy of the proposed approach is examined through two types of simulation studies and empirical analyses of the shadow economy and the fiscal illusion. [ABSTRACT FROM AUTHOR]