Real‐Time Fiscal Forecasting Using Mixed‐Frequency Data*

Date01 January 2020
Published date01 January 2020
DOIhttp://doi.org/10.1111/sjoe.12338
Scand. J. of Economics 122(1), 369–390, 2020
DOI: 10.1111/sjoe.12338
Real-Time Fiscal Forecasting Using
Mixed-Frequency Data*
Stylianos Asimakopoulos
University of Bath, Bath BA2 7AY, UK
s.asimakopoulos@bath.ac.uk
Joan Paredes
European Central Bank, DE-60640 Frankfurt, Germany
joan.paredes@ecb.europa.eu
Thomas Warmedinger
European Central Bank, DE-60640 Frankfurt, Germany
thomas.warmedinger@ecb.europa.eu
Abstract
The sovereign debt crisis has increased the importance of monitoring budgetary execution. We
employ real-time data using a mixed data sampling (MiDaS) methodology to demonstrate how
budgetary slippages can be detected early on. We show that in spite of using real-time data,
the year-end forecast errors diminish significantly when incorporating intra-annual information.
Our results show the benefits of forecasting aggregates via subcomponents, in this case total
government revenue and expenditure. Our methodology could significantly improve fiscal
surveillance and could therefore be an important part of the European Commission’s model
toolkit.
Keywords: Fiscal policy; mixed-frequency data; real-time data; short-term forecasting
JEL classification:C22; C53; E62; H68
I. Introduction
The sovereign debt crisis has highlighted the importance of fiscal
surveillance. In the context of the European Semester, it is indeed
important to assess the implications of incoming intra-annual data for
annual budgetary outturns. The usefulness of intra-annual fiscal data has
been shown in many recent studies, for example, P´erez (2007), Onorante
*We thank seminar participants at the European Central Bank (ECB) and at the European
Accounting Association Congress for helpful comments and suggestions. We are also grateful
to Eric Ghysels, Christian Schumacher, and Maria Dimou for providingsupport, comments, and
suggestions. The views expressed in this paper are those of the authors and do not necessarily
reflect those of the ECB. This paper has been circulated as an earlyversion entitled “Forecasting
fiscal time series using mixed frequency data”, ECB Working Paper Series, No. 1550.
C
The editors of The Scandinavian Journal of Economics 2018.
370 Real-time fiscal forecasting using mixed-frequency data
et al. (2010), Pedregal and P´erez (2010), and Paredes et al. (2014) using
data for the Euro Area, and Ghysels and Ozkan (2015) using data for the
United States. The focus of this existing body of literature is mainly on the
usefulness of high-frequency data to monitor the current-year government
balance (total revenue minus total expenditure). This is useful as such and
can be used to signal risks to budgetary executions. Hughes Hallett et al.
(2012) show how those signals should be used to design the necessary fiscal
corrections and the gains that can be achieved by such interventions.
The mixed data sampling (MiDaS) technique has been developed to
accurately project lower-frequency data with higher-frequency regressors.1
Using this technique, our aim in this paper is to utilize many disaggregated
real-time quarterly fiscal data to forecast annual data. We also examine the
differences between direct forecasts of aggregate fiscal variables and indirect
forecasts via their subcomponents, and we find that the latter works better.
MiDaS has been used in volatility predictions for financial sector data
(e.g., Ghysels et al., 2006; Forsberg and Ghysels, 2007) and in forecasting
macroeconomic variables using intra-annual data; see, for example, Bai
et al. (2013), Clements and Galvao (2008, 2009), and Kuzin et al. (2011),
who use monthly data to improve the quarterly forecast of macroeconomic
time series. Moreover, Asimakopoulos et al. (2017) have incorporated
MiDaS for the analysis of the predictability of dividend growth via a
time-disaggregated dividend–price ratio. Finally, Andreou et al. (2010)
and Ghysels and Wright (2009) use daily financial data to nowcast
macroeconomic data of monthly or quarterly frequency.
Our major advancement, compared with the papers in the field of intra-
annual fiscal data mentioned above, is that we use unrevised vintage data
(“real time”), which allows us to perform a real-time forecast.2This real-
time approach confirms the information content of quarterly fiscal data:
in particular, the year-end forecast errors diminish when incorporating
intra-annual information. Additionally, following Aruoba (2008), we present
some stylized facts of fiscal data revisions. Finally, following de Castro
et al. (2013), we also extend the analysis by incorporating macroeconomic
indicators to assess the rationality of those revisions.
1It should be noted that the purpose of this paper is not to compare models exhaustively in order
to find the best model for forecasting fiscal policy variables. For this reason, we do not assess
all possible approaches that can deal with mixed-frequency data. Nevertheless, we show that
MiDaS is a suitable tool for assessing mixed-frequency data in the context of fiscal policy. The
advantage of this approach, compared to alternative approaches such as state-space and mixed-
frequency vector autoregression models that make use of the Kalman filter,is that MiDaS is more
parsimonious and less sensitiveto specification er rors due to the use of non-linear lag polynomials
(e.g., Bai et al., 2013).
2In doing so, we avoid the potential issue of having misleading conclusions, as highlighted by
Orphanides (2001) and Cimadomo (2012).
C
The editors of The Scandinavian Journal of Economics 2018.

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT