Journal of the Japan Statistical Society, Japanese Issue
Online ISSN : 2189-1478
Print ISSN : 0389-5602
ISSN-L : 0389-5602
Volume 44, Issue 1
Displaying 1-14 of 14 articles from this issue
Article
  • Hiroyuki Takeuchi
    2014 Volume 44 Issue 1 Pages 1-17
    Published: September 26, 2014
    Released on J-STAGE: April 30, 2015
    JOURNAL FREE ACCESS
    If a non-degenerated probability distribution F has an analytic chracteristic function, there exists a strictly increasing function, so called a saddlepoint. The saddlepoint uniguely determines the distribution by its behavior only in a neighborhood of the expectation, and is being a line if and only if F is the normal (Takeuchi (2013)). In this paper we shall show that a local convexity of the saddlepoint includes much information of the corresponding distribution, with using sp-curvature which exists even for non-absolutely continuous distributions, and also for non-parametric case. It should be noted that the sp-curvature can determine the corresponding probability distribution uniquely, with moments up to second order, and it naturally explains the performance of the asymptotic normality of a statistics.
    Download PDF (830K)
Special Section: Economic Analysis and Structural Change
  • Ryo Kinoshita, Kosuke Oya
    2014 Volume 44 Issue 1 Pages 19-40
    Published: September 26, 2014
    Released on J-STAGE: April 30, 2015
    JOURNAL FREE ACCESS
    Structural change is gauged with the change of parameters in the model. In the case of multiple time series model, the causality between the time series also changes when there is a structural change. However the magnitude of change in causality is not clear in the case of structural change. We explore the measure of causality change between the time series and propose the test statistic whether there is any significance change in the causal relationship using frequency domain causality measure given by Geweke (1982) and Hosoya (1991). These procedures can be applied to error correction model which is non-stationary time series. The properties of the measure and test statistic are examined through the Monte Carlo simulation. As an example of application, the change in causality between United states and Japanese stock indexes is tested.
    Download PDF (1223K)
  • Makoto Takahashi
    2014 Volume 44 Issue 1 Pages 41-60
    Published: September 26, 2014
    Released on J-STAGE: April 30, 2015
    JOURNAL FREE ACCESS
    Realized Volatility (RV), which is computed as a squared sum of intraday returns, is a precise estimator of latent voaltility but is biased due to market microstructure effects. Takahashi et al. (2009) proposed a realized stochastic volatility (RSV), which models daily returns and RV simultaneously and adjusts the bias in RV\null. The RSV model assumes a constant mean of volatility despite we observe low and high volatilities in a boom-and-bust cycle. This article proposes a smooth transition RSV (STRSV), which models a time-varying mean of volatility by a smooth transition function, and shows a Bayesian estimation method via Markov chain Monte Carlo. Empirical analysis with the data of Japanese and the U.S. stock indices shows that the STRSV model captures the volatility dynamics appropriately and provides better fit to the data compared to the standard RSV model.
    Download PDF (1034K)
  • Daisuke Yamazaki, Eiji Kurozumi
    2014 Volume 44 Issue 1 Pages 61-74
    Published: September 26, 2014
    Released on J-STAGE: April 30, 2015
    JOURNAL FREE ACCESS
    This paper investigates tests for a shift in mean for serially correlated time series. It is known in the literature that LM-type tests suffer from the so-called non-monotonic power problem in the sense that the power decreases even if the magnitude of break gets larger. On the other hand, Wald-type tests do not have the non-monotonic power problem but they tend to suffer from over-size distortion. Several methods to overcome these problems have been developed in the literature and we propose a new test with a good finite sample property by modifying the existing test. We investigate the finite sample property of our new method and it turns out that our test performs better than the other tests in view of both size and power.
    Download PDF (849K)
  • Yohei Yamamoto
    2014 Volume 44 Issue 1 Pages 75-95
    Published: September 26, 2014
    Released on J-STAGE: April 30, 2015
    JOURNAL FREE ACCESS
    This paper statistically investigates structural change in the Japanese Phillips curve models and their forecasting performance stability. I find that there are three significant mean shifts in Japanese inflation data from 1970 to 2013. By considering them as exogenous effects, I did not find significant structural change in the Phillips curve slope coefficients in most cases of various real economic activity measures. However, if the model includes inflation expectation in the manner of rational expectation, the coefficients are found to be less stable. I also find evidence of so-called forecast beakdown in inflation forecasts if I use the Phillips curve model estimated by the data after 1982 in which the coefficients are deemed to be stable, together with inflation data with the mean shifts to account for possible exogenous mean shifts in the future.
    Download PDF (2023K)
Special Section: Covariate Shift and Density Ratio Estimation
  • Takafumi Kanamori
    2014 Volume 44 Issue 1 Pages 97-111
    Published: September 26, 2014
    Released on J-STAGE: April 30, 2015
    JOURNAL FREE ACCESS
    A density ratio is defined by the ratio of two probability densities, and it is widely applied to real world statistical problems. In this paper, we introduce moment matching estimators of density ratios and f-divergence estimators based on density-ratios. Then, we exploit the f-divergence estimators to the two-sample homogeneity test. We prove that the proposed test dominates the existing empirical likelihood score test. Through numerical studies, we illustrate the adequacy of the asymptotic theory for finite-sample inference.
    Download PDF (790K)
  • Masashi Sugiyama, Makoto Yamada, Marthinus Christoffel du Plessis, Son ...
    2014 Volume 44 Issue 1 Pages 113-136
    Published: September 26, 2014
    Released on J-STAGE: April 30, 2015
    JOURNAL FREE ACCESS
    In standard supervised learning algorithms training and test data are assumed to follow the same probability distribution. However, because of a sample selection bias or non-stationarity of the environment, this important assumption is often violated in practice, which causes a significant estimation bias. In this article, we review semi-supervised adaptation techniques for coping with such distribution changes. We focus on two scenarios of such distribution change: the covariate shift (input distributions change but the input-output dependency does not change) and the class-balance change in classification (class-prior probabilities change but class-wise input distributions remain unchanged). We also show methods of change detection in probability distributions.
    Download PDF (1272K)
Special Section: Developments and Future Issues in Macroeconometric Time Series Analysis
  • Tatsuyoshi Okimoto
    2014 Volume 44 Issue 1 Pages 137-157
    Published: September 26, 2014
    Released on J-STAGE: April 30, 2015
    JOURNAL FREE ACCESS
    Many economic and financial data seem to change their behavior depending on the business cycle and/or policy regime. In this paper, we review the Markov switching (MS) model as one of the most powerful tools to analyze such economic and financial data with switching regimes. More specifically, following the brief introduction of the MS model, we discuss the Markov chain which is an important component of the model and explain how to interpret the MS model using a simple example. Lastly, we argue the statistical inference associated with the MS model and provide some applications to macroeconomics and finance.
    Download PDF (757K)
  • Takashi Kano
    2014 Volume 44 Issue 1 Pages 159-187
    Published: September 26, 2014
    Released on J-STAGE: April 30, 2015
    JOURNAL FREE ACCESS
    This paper critically reviews roles of dynamics stochastic general equilibrium (DSGE) models in macroeconometrics, introducing econometric categorizations of DSGE models made by Geweke (2010): strong, weak, and minimal econometric interpretations. As an application of the minimal interpretation, this paper introduces the Bayesian Monte Carlo exercise conducted by Kano and Nason (2014) for investigating business cycle implications of consumption habits as a propagation mechanism of monetary policy shocks.
    Download PDF (1119K)
  • Koiti Yano
    2014 Volume 44 Issue 1 Pages 189-216
    Published: September 26, 2014
    Released on J-STAGE: April 30, 2015
    JOURNAL FREE ACCESS
    Particle filters and smoothers are simulation-based methods to estimate non-linear non-Gaussian state space models. The filters and smoothers are widely applied to science and engineering from the early 1990s. We describe an introduction to the particle filter and some applications in Section 2. The particle fixed-lag smoother is denoted in Section 3, and we apply the resample-move method to the particle fixed-lag smoother in Section 4. We explain parameter estimation and a self-organzing state space model in Section 5. In Section 6, we estimate a Real Business Cycle model based on the filter.
    Download PDF (1093K)
Special Topic: The JSS Prize Lecture
Book reviews
feedback
Top