Temporal disaggregation by state space methods: Dynamic regression methods revisited

Proietti, Tommaso

doi:10.1111/j.1368-423x.2006.00189.x

Cited by 59 publications

(58 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…However, the empirical and Monte Carlo evidence show that its performance is sometimes disappointing. This is due to the flatness of the implied likelihood profile and, therefore, the corresponding observational equivalence in a wide range of values for its dynamical parameter, see Proietti (2006). On the other hand, the methods of Santos Silva-Cardoso and Proietti place the dynamics in the variables y and x, treating the innovation as a purely random shock.…”

Section: Initial Quarterly Regional Gdp Estimationmentioning

confidence: 99%

“…We have considered several benchmarking procedures for deriving the preliminary GDP estimates: Chow andLin (1971), Fernández (1981), Santos-Silva and Cardoso (2001) and Proietti (2006). All of them hinge upon a dynamic linear model that links the (observable) high-frequency indicator with the (unobservable) regional GDP.…”

Section: Initial Quarterly Regional Gdp Estimationmentioning

confidence: 99%

“…Preliminary estimates of quarterly GDP at the regional level are compiled using benchmarking techniques (see Di Fonzo 1987, 2002and Proietti 2006 for an in-depth exposition). These techniques play an important role in the compilation practices of QNA around the world (see Eurostat 1998 andBloem et al 2001).…”

Section: Initial Quarterly Regional Gdp Estimationmentioning

confidence: 99%

See 2 more Smart Citations

Quarterly Regional GDP Flash Estimates by Means of Benchmarking and Chain Linking

Cuevas¹,

Quilis²,

Espasa

2015

Journal of Official Statistics

View full text Add to dashboard Cite

In this article we propose a methodology for estimating the GDP of a country's different regions, providing quarterly profiles for the annual official observed data. Thus the article offers a new instrument for short-term monitoring that allows the analysts to quantify the degree of synchronicity among regional business cycles. Technically, we combine time-series models with benchmarking methods to process short-term quarterly indicators and to estimate quarterly regional GDPs ensuring their temporal and transversal consistency with the National Accounts data. The methodology addresses the issue of nonadditivity, explicitly taking into account the transversal constraints imposed by the chain-linked volume indexes used by the National Accounts, and provides an efficient combination of structural as well as short-term information. The methodology is illustrated by an application to the Spanish economy, providing real-time quarterly GDP estimates, that is, with a minimum compilation delay with respect to the national quarterly GDP. The estimated quarterly data are used to assess the existence of cycles shared among the Spanish regions.

show abstract

Section: Initial Quarterly Regional Gdp Estimationmentioning

confidence: 99%

Section: Initial Quarterly Regional Gdp Estimationmentioning

confidence: 99%

See 1 more Smart Citation

Quarterly Regional GDP Flash Estimates by Means of Benchmarking and Chain Linking

Cuevas¹,

Quilis²,

Espasa

2015

Journal of Official Statistics

View full text Add to dashboard Cite

show abstract

“…This is achieved in practice using the state space methods set up in Harvey and Chun (2000) and Proietti (2006). In order to handle temporal aggregation, a new state space representation is derived from that of the underlying true model, an IMA(1,1) model, by augmenting the state vector of the original state space representation with a cumulator variable that is only partially observed.…”

Section: An Artificial Examplementioning

confidence: 99%

Band Spectral Estimation for Signal Extraction

Proietti

2007

SSRN Journal

Self Cite

View full text Add to dashboard Cite

The paper evaluates the potential of band spectral estimation for extracting signals in economic time series. Two situations are considered. The first deals with trend extraction when the original data have been permanently altered by routine operations, such as prefiltering, temporal aggregation and disaggregation, and seasonal adjustment, which modify the high frequencies properties of economic time series. The second is when the measurement model is only partially specified, in that it aims at fitting the series in a particular frequency range, e.g. at interpreting the long run behaviour. These issues are illustrated with reference to a simple structural model, namely the random walk plus noise model.

show abstract

“…Further improvements can be obtained, for instance, by assuming that residuals follow a random walk (Fernández, 1981). Apart from changing assumptions on the underlying autocorrelation structure of the residuals, previous research improved estimates by adapting the Chow and Lin (1971) methodology to various settings: Santos Silva and Cardoso (2001) to dynamic models and Proietti (2006) to state space models, among others.…”

Section: Introductionmentioning

confidence: 99%

Chow-Lin×N: How adding a panel dimension can improve accuracy

Bettendorf

Bursian

2017

Economics Letters

View full text Add to dashboard Cite

Practitioners and academics are often faced with the problem that some time series of interest are not available at high frequencies. When related series exist, the Chow and Lin (1971) methodology can be used to disaggregate a low frequency series into its high frequency counterpart using the available related series. The aim of this paper is to increase accuracy of the Chow and Lin (1971) methodology by exploiting information from the cross-sectional dimension. ContributionWe suggest jointly estimating multiple Chow and Lin (1971) equations, one for each crosssectional unit, restricting the coefficients to be the same across units in order to interpolate unit-specific data. In contrast to the estimation of single equations, unobservable timevarying characteristics (e.g. structural breaks) that are common across units can be taken into account. The proposed approach is straightforward to implement and can readily be applied to various settings. ResultsWe provide empirical evidence that the panel-based approach can improve accuracy compared to single equation models, in particular when the time dimension of available data is short. Furthermore, the results suggest that controlling for unobservable time-varying characteristics can improve accuracy of the resulting interpolated series. Abstract Single equation models are well established among academics and practitioners to perform temporal disaggregation of low frequency time series using available related series. In this paper, we propose an extension that exploits information from the cross-sectional dimension. More specifically, we suggest jointly estimating multiple Chow and Lin (1971) equations, one for each cross-sectional unit (e.g. country), restricting the coefficients to be the same across units in order to interpolate unitspecific data. Using actual data on real GDP and industrial production for euro area countries we provide evidence that this approach can result in more accurate interpolated time series for individual countries. The results suggest that the inclusion of time fixed effects, which is not feasible in standard single equation models, can be helpful in increasing accuracy of the resulting series.

show abstract

Temporal disaggregation by state space methods: Dynamic regression methods revisited

Cited by 59 publications

References 21 publications

Quarterly Regional GDP Flash Estimates by Means of Benchmarking and Chain Linking

Quarterly Regional GDP Flash Estimates by Means of Benchmarking and Chain Linking

Band Spectral Estimation for Signal Extraction

Chow-Lin×N: How adding a panel dimension can improve accuracy

Contact Info

Product

Resources

About