Interpolation and backdating with a large information set

Abstract

Reproduction for educational and non-commercial purposes is permitted provided that the source is acknowledged.

Estimating and forecasting the euro area monthly national accounts from a dynamic factor model

“…Note that these estimates are obtained from the Kalman smoother based on the entire data set Z T . Angelini et al (2005) compare factor-based interpolation methods with the traditional method by Chow and Lin (1971) and conclude that both methods fare well. We therefore also inspect estimates of monthly GDP growth from applying the Chow-Lin method to a single equation.…”

A comparison of methods for the construction of composite coincident and leading indexes for the UK

“…We will refer to the non-parametric factor based indicators as SW2 and FHLR2 CCIs. 2 In the three panels of Figure 1 we graph the levels, six month percentage changes and filtered versions of the (standardized) NMB CCI and of the three versions of the factor based CCIs, namely, SW, SW2 and FHLR2. 3 The different CCIs seem to move component included into the VAR, and the type of cointegration test applied (Johansen's (1988) trace or eigenvalue statistic).…”

“…Since all the factor based methods require the input variables to be weakly stationary, we model the log differences of the single coincident indicators. 2 More specifically, the SW2 CCI is the first static principal component of the four coincident series, while to construct the FHLR2 CCI we apply their two-step procedure, set the bandwidth parameter at M=12, and use one factor both in the first step (i.e. to compute the variance covariance matrix of the common components obtained using FHLR) and in the second step.…”

Econometric analyses with backdated data: Unified Germany and the euro area

“…However, from an econometric point of view, an improper treatment of the problem can introduce a substantial measurement error, which in turn can bias all the results of the analysis. 1 In this paper we propose to apply more sophisticated techniques to backdate German data prior to the re-unification. Specifically, based on the empirical results in Angelini, Henry and Marcellino The idea underlying both methods is to regress the series of interest, which contains missing observations at the beginning of the sample, on a set of series covering the whole sample.…”

“…When the set of potentially significant explanatory variables is large (compared with the sample 1 The construction of euro area time series presents other problems, not considered in this paper to focus on the main issue of German unification, such as the proper choice of a weighting scheme or the treatment of seasonality, see We have also considered univariate and multivariate time series models (AR and VAR) for the unified German series and applied the Kalman smoother to backdate the missing observations. This is basically equivalent to reverting the order of the observations in the time series and compute dynamic forecasts for them with a forecast horizon from h=1 to h=84 (to recover quarterly data in the '70s and '80s).…”