Fiscal Foresight, Limited Information and the Effects of Government Spending Shocks
| DOI | http://doi.org/10.1111/obes.12036 |
| Author | Emanuel Gasteiger,Matteo Fragetta |
| Date | 01 October 2014 |
| Published date | 01 October 2014 |
667
©2013 The Department of Economics, University of Oxford and JohnWiley & Sons Ltd.
OXFORD BULLETIN OF ECONOMICSAND STATISTICS, 76, 5 (2014) 0305–9049
doi: 10.1111/obes.12036
Fiscal Foresight, Limited Information and the Effects
of Government Spending Shocks*
Matteo Fragetta†,‡ and Emanuel Gasteiger§
†University of Salerno, Department of Economics and Statistics, Via Ponte Don Melillo,
84084, Fisciano (SA) Italy (e-mail: mfragetta@unisa.it)
‡Instituto Universit´ario de Lisboa (ISCTE – IUL) BRU – IUL, Lisboa, Portugal
§Instituto Universit´ario de Lisboa (ISCTE – IUL), BRU – IUL, Department of Economics,
Av.adas For¸cas Armadas, 1649-026, Lisboa, Portugal (e-mail: emanuel.gasteiger@iscte.pt)
Abstract
We quantify the impact of government spending shocks in the US. Thereby, wecontrol for
fiscal foresight, a specific limited information problem (LIP) by utilizing the narrative
approach. Moreover, we surmount the generic LIP inherent in vector autoregressions
(VARs) by a factor-augmented VAR (FAVAR) approach. We find that a positive deficit-
financed defence shock raises output by more than in a VAR (e.g. 2.61 vs. 2.04 for peak
multipliers). Furthermore, our evidence suggests that consumption is crowded in. These
results are robust to variants of controlling for fiscal foresight and reveal the crucial role
of the LIP in fiscal VARs.
I. Introduction
Differing empirical findings on the effect of government spending shocks on key variables
such as consumption and real wage amongst others depend ultimately on the empirical
approach employed.One consequence of different empirical findings is a controversy about
the most suitablemodel to analyze fiscal policy in theory. Perotti(2007, p. 1) gets to the point:
…perfectly reasonable economists can and do disagree on the basic theoretical effects of
fiscal policy, and on the interpretation of the existing empirical evidence.
With regard to empirical findings, there exists the classic structural VAR (SVAR)
approach employed in Fat´as and Mihov (2001), Blanchard and Perotti (2002), Perotti
(2007) and Fragetta and Melina (2011) amongst others. These articles find an increase of
consumption and real wage in response to a government spending shock in the US, which
supports (new-)Keynesian theories.
*Weare indebted to an anonymous referee, Sergio Destefanis, Robert Kunst, Lu´ıs Martins, Giovanni Melina, Lucio
Sarno and the participants of the DIW Berlin Macroeconometric Workshop2011, the 3rd International Conference in
memory of Prof. Carlo Giannini at the Banca D’Italia, the 6th Meeting of the Portuguese Economic Journal and the
27th Annual Congress of the European Economic Association/66th European Meeting of the Econometric Society
for many helpful comments. Furthermore, we are grateful to Studio Lembo for enduring hospitality. Gasteiger
acknowledges financial support from Funda¸c˜ao para a Ciˆencia e a Tecnologia (PEst-OE/EGE/UI0315/2011). All
remaining errors are the responsibility of the authors.
JEL Classification numbers: C32, E21, E32, E62, H30, H50.
668 Bulletin
An alternative is the narrative approach followed by Ramey and Shapiro (1998), Edel-
berg, Eichenbaum and Fisher (1999), Burnside, Eichenbaum and Fisher (2004) and Ramey
(2011b) which finds that consumption and real wage decrease in response to a government
spending shock in the US. These results back neoclassical theories. The virtue of these
articles is that they overcome the fiscal foresight problem related to anticipation effects by
economic agents that cause a misalignment of information sets.1
Note that, although prominently treated in the fiscal vector autoregression (VAR)
literature, fiscal foresight is just one specific manifestation of how the information set
of an econometrician on the one side and the economy he aims to analyze on the other side
can become misaligned. Fiscal foresight is essentially a limited information problem.The
latter can emerge due to other generic causes, as we discuss below.
The term fiscal foresight captures the fact that a pre-announced fiscal policy change
leads agents in the actual economy to revise their plans even before the policy change
becomes effective. In consequence, the information set of the agents in the actual economy
is larger than the information set of the econometrician that aims to analyze the data
set generated by the actual economy. In case the econometrician makes use of a classic
SVAR approach, the differing information sets might flaw statistical inferences. What the
econometrician beliefs to be a fiscal innovation is actually a discounted sum of current and
past fiscal news observed by agents.2Thus, one could argue that conclusions drawn from
analyzes following the SVAR approach must be viewed cautiously.
In contrast, the narrative approach pioneered by Hamilton (1985) involves an identifi-
cation strategy that adequately accounts for fiscal foresight and helps to align the informa-
tion set of the agents in the actual economy and the econometrician. The basic idea of this
approach nowadays is to build time series that contain the net present value of announced
government spending or tax code changes that are going to be effective at a later date and
that are not related to the state of the economy.3Such a time series is unmistakably fiscal
news to both, the agents in the actual economy and the econometrician. We regard the nar-
rative approach as a consistent and convincing identification strategy in order to properly
account for fiscal foresight. Thus, we simply follow this approach in our analysis below.
Nevertheless, fiscal foresight is not the only source that can misalign the information set
of the econometrician and the actual economy.
Our main goal is to highlight that misalignments in information sets, that may lead to
biased estimates or non-fundamentalness, can emerge generically in fiscal VARs and need
to be addressed.
1Ramey (2011a) highlights the key differences in neoclassical and (new-)Keynesiantheories and sur veysrecent
empirical findings based on the SVARand narrative approach in the light of these theories.
2Leeper, Walker andYang (2009), Mertens and Ravn (2010) or Favero and Giavazzi (2012) illustrate the fiscal
foresight problem in more detail.
3Originally the idea was to conduct an event-study, where the events are not related to the state of the economy.
Ramey and Shapiro (1998) have initiated a different, narrative, dummy variable approach to identify shocks to
government spending focusing only on episodes where sudden forecasts of large rises in defence spending due to
exogenous events where announced bythe Business Week magazine. More recently Ramey (2011b), in order to get
more information, has quantified the military news, building a continuous variablewhich contains the discounted value
of the resulting change in government spending forecasted by the BusinessWeek. With this ‘defence news’measure,
Ramey (2011b) overallconfir ms the previousfinding arguing that SVAR analysis reach different conclusion because
they miss the right timing.
©2013 The Department of Economics, University of Oxford and JohnWiley & Sons Ltd
To continue reading
Request your trial