Minimax designs for approximately linear models with AR(1) errors

We obtain designs for linear regression models under two main departures from the classical assumptions: (1) the response is taken to be only approximately linear, and (2) the errors are not assumed to be independent, but to instead follow a first-order autoregressive process. These designs have the property that they minimize (a modification of) the maximum integrated mean squared error of the estimated response, with the maximum taken over a class of departures from strict linearity and over all autoregression parameters ρ,|ρ,| < 1, of fixed sign. Specific methods of implementation are discussed. We find that an asymptotically optimal procedure for AR(1) models consists of choosing points from that design measure which is optimal for uncorrelated errors, and then implementing them in an appropriate order.     

Tree-structured generalized autoregressive conditional heteroscedastic models

We propose a new generalized autoregressive conditional heteroscedastic (GARCH) model with tree-structured multiple thresholds for the estimation of volatility in financial time series. The approach relies on the idea of a binary tree where every terminal node parameterizes a (local) GARCH model for a partition cell of the predictor space. The fitting of such trees is constructed within the likelihood framework for non-Gaussian observations: it is very different from the well-known regression tree procedure which is based on residual sums of squares. Our strategy includes the classical GARCH model as a special case and allows us to increase model complexity in a systematic and flexible way. We derive a consistency result and conclude from simulation and real data analysis that the new method has better predictive potential than other approaches.     

CANONICAL VARIATE ANALYSIS WITH SPATIALLY-CORRELATED DATA

A general canonical variate model is derived when the observations are spatially correlated. For spatial covariance structures resulting from dependence of a pixel on its nearest neighbours, the solution reduces to an analysis of neighbour-corrected values. The usual analysis, in which spatial correlation is ignored, gives similar canonical vectors but over-estimates the canonical roots. A formula for approximating the reduction in the canonical roots to adjust for the spatial correlation is given.     

Monetary policy,trade-offs and the transmission of UK Monetary Policy

We provide a new method to model changes in monetary policy of the Bank of England (BoE) as well as the impact of these changes on UK economy. This is important as central bankers have widened the range of instruments in their monetary policy toolbox. Specifically, we estimate a proxy for the monetary policy stance and then analyse a time-varying parameter vector autoregressive with stochastic volatility model to explain the BoE’s trade-offs when making policy decisions and as well as to demonstrate dynamic impacts of monetary policy on inflation and economic growth. The empirical results show that our estimated monetary policy proxy is better at capturing the BoE’s policy when the interest rate lower bound becomes binding.     

Problems related to confidence intervals for impulse responses of autoregressive processes

Confidence intervals for impulse responses computed from autoregressive processes are considered. A detailed analysis of the methods in current use shows that they are not very reliable in some cases. In particular, there are theoretical reasons for them to have actual coverage probabilities which deviate considerably from the nominal level in some situations of practical importance. For a simple case alternative bootstrap methods are proposed which provide correct results asymptotically.     

Spatial smoothing of low birth weight rate in Bangladesh using Bayesian hierarchical model

The term low birth weight refers an event where a newborn baby has a weight that is less than 2500?g. This is an essential indicator while the interest is in public health issues such as infant mortality, maternal complications, and antenatal care, etc. of a country, particularly, for a developing country like Bangladesh. The regional development programs are in the current priority list of Bangladesh government and other policy makers. Many of such regional development programs may need the spatial distribution of relative risk for low birth weight that can be obtained by mapping the risks over small area domains like the districts of Bangladesh. This study aims to find whether is there any spatial dependence among the relative risks of low birth weight for the districts of Bangladesh. This has been investigated using Moran's I statistic and a significant spatial dependence in the relative risks was found. Then, attempt has been made to rediscover the spatial distribution based on the idea of spatial smoothing. A Bayesian hierarchical model is used considering percent received antenatal care and female labor force participation as covariates to smooth the observed relative risks of low birth weight in 64 districts of Bangladesh. Revised spatial distribution taking the spatial dependence under consideration through intrinsic conditional autoregressive model is derived and showed in choropleth map along with its different behaviors.     

A NOTE ON THE INVERSE COVARIANCE MATRIX OF THE AUTOREGRESSIVE PROCESS1

The inverse covariance matrix of an autoregressive process of arbitrary order is found explicitly using the notion of the information matrix for the vector random variable, not the parameters. Any process for which a simple conditional representation exists, can be treated in the same way. The approach has merit in the teaching of statistics.     

THE SIGN TEST FOR STOCHASTIC PROCESSES

It is shown that the sign test may be applied to the residuals from the use of model fitting procedures, such as conditional least squares, to make inferences on the predictable part of a stochastic process. Minimal assumptions on the distribution of the process, apart from those already required for the model fitting procedure, are needed. The results are illustrated with an application to first order autoregressive processes.     

Likelihood Analysis of the I(2) Model

The I (2) model is defined as a submodel of the general vector autoregressive model, by two reduced rank conditions. The model describes stochastic processes with stationary second difference. A parametrization is suggested which makes likelihood inference feasible. Consistency of the maximum likelihood estimator is proved, and the asymptotic distribution of the maximum likelihood estimator is given. It is shown that the asymptotic distribution is either Gaussian, mixed Gaussian or, in some cases, even more complicated.     

Functional Coefficient Regression Models for Non-linear Time Series: A Polynomial Spline Approach

Abstract.  We propose a global smoothing method based on polynomial splines for the estimation of functional coefficient regression models for non-linear time series. Consistency and rate of convergence results are given to support the proposed estimation method. Methods for automatic selection of the threshold variable and significant variables (or lags) are discussed. The estimated model is used to produce multi-step-ahead forecasts, including interval forecasts and density forecasts. The methodology is illustrated by simulations and two real data examples.