Showing papers by "Serena Ng published in 2002"

Determining the Number of Factors in Approximate Factor Models

Abstract: In this paper we develop some econometric theory for factor models of large dimensions. The focus is the determination of the number of factors (r), which is an unresolved issue in the rapidly growing literature on multifactor models. We first establish the convergence rate for the factor estimates that will allow for consistent estimation of r. We then propose some panel criteria and show that the number of factors can be consistently estimated using the criteria. The theory is developed under the framework of large cross-sections (N) and large time dimensions (T). No restriction is imposed on the relation between N and T. Simulations show that the proposed criteria have good finite sample properties in many configurations of the panel data encountered in practice.

Determining the Number of Factors in Approximate Factor Models

Abstract: In this paper we develop some econometric theory for factor models of large dimensions. The focus is the determination of the number of factors (r), which is an unresolved issue in the rapidly growing literature on multifactor models. We first establish the convergence rate for the factor estimates that will allow for consistent estimation of r. We then propose some panel criteria and show that the number of factors can be consistently estimated using the criteria. The theory is developed under the framework of large cross-sections (N) and large time dimensions (T). No restriction is imposed on the relation between N and T. Simulations show that the proposed criteria have good finite sample properties in many configurations of the panel data encountered in practice.

PPP May not Hold Afterall: A Further Investigation

Abstract: *Department of EconomicsBoston UniversityE-mail: perron@bu.eduIn a recent paper, Engel, C. (1999) presents monte-carlo evidence to suggestthat unit root tests cannot detect a non-stationary component in the realexchange rate even when this component accounts for almost half of its long-horizon forecast error variance. This hidden non-stationary component led tothe conclusion that long run purchasing power parity might not hold afterall.In this note, we first point out some conceptual difficulties with the statisticbeing used to measure the size of the non-stationary component, and thenargue that it bears no systematic relationship with rejection rates in unit roottests. The problems stem from near observational equivalence of the simulatedmodel in not one, but two dimensions. We then discuss the steps a practitionercan take to minimize Type I error in cases when the non-stationary componentis hard to detect. Real exchange rate data for 19 countries are examinedand estimates are obtained for the duration of the real exchange rate shocks.

Analysis of vector autoregressions in the presence of shifts in mean

Abstract: This paper considers the implications of mean shifts in a multivariate setting. It is shown that under the additive outlier type mean shift specification, the intercept in each equation of the vector autoregression (VAR) will be subject to multiple shifts when the break dates of the mean shifts to the univariate series do not coincide. Conversely, under the innovative outlier type mean shift specification, both the univariate and the multivariate time series are subject to multiple shifts when mean shifts to the innovation processes occur at different dates. We consider two procedures, the first removes the shifts series by series before forming the VAR, and the second removes intercept shifts in the VAR directly. The pros and cons of both methods are discussed.

PPP May not Hold Afterall: A Further Investigation

Abstract: In a recent paper, Engel, C. (1999) presents monte-carlo evidence to suggest that unit root tests cannot detect a non-stationary component in the real exchange rate even when this component accounts for almost half of its longhorizon forecast error variance. This hidden non-stationary component led to the conclusion that long run purchasing power parity might not hold afterall. In this note, we first point out some conceptual difficulties with the statistic being used to measure the size of the non-stationary component, and then argue that it bears no systematic relationship with rejection rates in unit root tests. The problems stem from near observational equivalence of the simulated model in not one, but two dimensions. We then discuss the steps a practitioner can take to minimize Type I error in cases when the non-stationary component is hard to detect. Real exchange rate data for 19 countries are examined and estimates are obtained for the duration of the real exchange rate shocks.

Can Sticky Prices Account for the Variations and Persistence in Real Exchange Rates

Abstract: This paper provides an empirical assessment of the importance of sticky prices in accounting for the variations and the persistence in real exchange rates. Vector autoregressions with five variables from two countries that always include the United States are estimated. Restrictions are imposed to identify a global shock, and two sets of country specific output shocks. One set of shocks is associated with instantaneous price adjustments, while the other has delayed effects on prices. Data from the G7 countries reveal that U.S sticky price shocks are the dominant source of real exchange rate variations. But these shocks have reasonably short half-lives and cannot account for the observed real exchange rate persistence. Non-sticky price shocks can induce very persistent real exchange rate dynamics, even though they account for little of the historical real exchange rate variations.

Forecasting autoregressive time series in the presence of deterministic components

Abstract: This paper studies the error in forecasting an autoregressive process with a deterministic component. We show that when the data are strongly serially correlated, forecasts based on a model that detrends the data using OLS before estimating the autoregressive parameters are much less precise than those based on an autoregression that includes the deterministic components, and the asymptotic distribution of the forecast errors under the two-step procedure exhibits bimodality. We explore the conditions under which feasible GLS trend estimation can lead to forecast error reduction. The finite sample properties of OLS and feasible GLS forecasts are compared with forecasts based on unit root pretesting. The procedures are applied to 15 macroeconomic time series to obtain real time forecasts. Forecasts based on feasible GLS trend estimation tend to be more efficient than forecasts based on OLS trend estimation. A new finding is when a unit root pretest rejects non-stationarity, use of GLS yields smaller forecast errors than OLS. When the series to be forecasted is highly persistent, GLS trend estimation in conjunction with unit root pretests can lead to sharp reduction in forecast errors.

Demand Systems With Nonstationary Prices

Abstract: Relative prices are nonstationary and standard root-T inference is invalid for demand systems. But demand systems are nonlinear functions of relative prices, and standard methods for dealing with nonstationarity in linear models cannot be used. Demand system residuals are also frequently found to be highly persistent, further complicating estimation and inference. We propose a variant of the Translog demand system, the NTLOG, and an associated estimator that can be applied in the presence of nonstationary prices with possibly nonstationary errors. The errors in the NTLOG can be interpreted as random utility parameters. The estimates have classical root-T limiting distributions. We also propose an explanation for the observed nonstationarity of aggregate demand errors, based on aggregation of consumers with heterogeneous preferences in a slowly changing population. Estimates using US data are provided.

Forecasting Autoregressive Time Series in the Presence of Deterministic Components

Abstract: This paper studies the error in forecasting an autoregressive process with a deterministic component. We show that when the data are strongly serially correlated, forecasts based on a model that detrends the data using OLS before estimating the autoregressive parameters are much less precise than those based on an autoregression that includes the deterministic components, and the asymptotic distribution of the forecast errors under the two-step procedure exhibits bimodality. We explore the conditions under which feasible GLS trend estimation can lead to forecast error reduction. The finite sample properties of OLS and feasible GLS forecasts are compared with forecasts based on unit root pretesting. The procedures are applied to 15 macroeconomic time series to obtain real time forecasts. Forecasts based on feasible GLS trend estimation tend to be more efficient than forecasts based on OLS trend estimation. A new finding is when a unit root pretest rejects non-stationarity, use of GLS yields smaller forecast errors than OLS. When the series to be forecasted is highly persistent, GLS trend estimation in conjunction with unit root pretests can lead to sharp reduction in forecast errors.