Sectoral Effects of News Shocks
| Date | 01 April 2019 |
| Published date | 01 April 2019 |
| DOI | http://doi.org/10.1111/obes.12269 |
| Author | Marija Vukotić |
215
©2018 The Department of Economics, University of Oxford and JohnWiley & Sons Ltd.
OXFORD BULLETIN OF ECONOMICSAND STATISTICS, 81, 2 (2019) 0305–9049
doi: 10.1111/obes.12269
Sectoral Effects of News Shocks*
Marija Vukoti´
c†
†Department of Economics, University of Warwick, Gibbet Hill Road, Coventry CV47AL,
UK (e-mail: m.vukotic@warwick.ac.uk)
Abstract
This paper argues that an aggregate news shock reveals news about technological im-
provements in the durable goods sector. Better technological prospects translate into large
responses of the fundamentals in the durable goods sector; much larger than the responses
of the fundamentals in the non-durable goods sector.These better technological prospects,
contrary to common belief, do not induce short-run comovement among fundamentals
within either of the two sectors. The behaviour of inventories, an important margin that
durable goods producers can use to buffer news shocks, proves to be crucial for reconciling
the effects of news shocks in a two-sector model with the data.
I. Introduction
After being abandoned for more than half a century, the idea that the expectations about
future changes in productivity represent an important driving force of the business cycle
has experienced a revival, receiving a great deal of attention in the recent literature.1
Beaudry and Portier (2004, 2006) were the first authors to reassess the importance of news
about future technological developments as driversof business cycles. They find that news
shocks account for more than half of output (business-cycle) fluctuations and also induce
comovement among aggregate variables.2
The purpose of this paper is to gain deeper understanding about the nature of this
important shock by looking at the channels through which it propagates the business cycle.
Specifically, I analyse the behaviour of the manufacturing sector, because it allows for a
clear distinction between the non-durable and durable goods industries. As I will show,
the aggregate news shock is essentially a durable-goods-sector news shock, whichimplies
that durable goods industries play a dominant role in the propagation of an aggregate news
JEL Classification numbers: E3, E32, L60.
*I would like to thank the editor and two anonymous referees whose valuable suggestions improved the paper.
I would also like to thank Craig Burnside for his guidance during the early stage of this project, as well as Robert
Barsky, Francesco Bianchi, Christian Hellwig,Cosmin Ilut, Rober to Pancrazi,Pietro Peretto, Franck Portier, Barbara
Rossi, Juan Rubio-Ramirez, and Thijs Van Rens for their comments.
1Pigou (1927) was one of the first authors to propose that agents’ expectations about the future are an important
source of business-cycle fluctuations.
2Jaimovich and Rebelo (2009), Beaudry and Lucke (2010) and Schmitt-Groh´e and Uribe (2012) also find news
shocks to be an important driver of business-cycle fluctuations. Fora very detailed survey of the papers that contribute
to this literature, see Beaudry and Portier (2014).
216 Bulletin
shock. This result is consistent with Mankiw (1985), who concludes that durable goods
industries play an essential role in the business cycle, and that explaining fluctuations in
the durable goods sector is vital for understanding aggregate economic fluctuations.
This paper argues that an aggregate news shock is effectively a news shock about techno-
logical improvements in the durable goods sector. Better technological prospects translate
into large responses of the fundamentals in the durable goods sector; much larger than the
responses of the fundamentals in the non-durable goods sector.3These better technological
prospects, however, do not induce short-run comovement among fundamentals within ei-
ther of the two sectors. This lack of the short-run comovementcan be better understood by
looking at the behaviour of inventories, an important margin that durable goods producers
can use to buffer news shocks. In fact, my investigation of inventories within a two-sector
model proves to be crucial for understanding the propagation channel of an aggregate news
shock to the two sectors, and to reconciling the short-run effects of news shocks in a model
with the data.
My empirical analysis relies on the identification proposed by Barsky and Sims (2011),
in which an aggregate news shock is identified as the shock that has no contemporaneous
impact on total factor productivity (TFP) and that simultaneously explains most of its
forecast error variance overthe 10-year horizon.4The contribution of this empirical analysis
consists of two parts. First, my sector-focused investigation shows that an aggregate news
shock manifests as a durable-goods-sector newsshock, and, therefore, propagates primarily
through the durable goods sector. In particular, after a 1% aggregate news shock, the
response of durable-goods-sector productivity after a three-year horizon is already about
three times greater than the response of the non-durable-goods-sector productivity. This
higher productivity increase translates into significantly higher percentage responses of
fundamentals in the durable goods sector than in the non-durable goods sector. Second,
my sector-focused investigation also shows that a positive aggregate news shock does not
generate comovement among sectoral fundamentals within the two sectors. In particular,
a positive aggregate news shock leads to the following responses: positive investment in
both sectors; negative hours and output in both sectors. In addition, aggregate news shocks
introduce negative correlation in consumption across sectors, different from the positive
unconditional correlation observed in the data.5
It has been long understood that the producers of durables can stock inventories and
use them to buffer shocks. Nearly a century ago, Pigou (1927) proposed that the possi-
bility of holding stocks of inventories explained the fact that business-cycle fluctuations
3Throughout this paper, I consider fundamentals to be the followingvariables of interest: productivity, stock prices,
output, consumption, investment, and hours worked.The term ‘sectoral fundamentals’ refers to these variables in the
durable goods or non-durable goods sectors.
4My results are robust to the use of different time horizons and a slightly different identification, proposed by
Francis et al. (2014), and used by Beaudry, Nam and Wang(2011) for the purposes of identifying news shocks. My
results are also in line with Theodoridis and Zanetti (2016) who show that their findings are robust across different
identification strategies as well as across horizons of 40, 80 and 120 quarters.
5Long and Plosser (1983) were the first to point out that the comovementof different sectors represents an impor-
tant feature of business cycles. In addition, Christiano and Fitzgerald (1998) document a strong comovementbetween
hours worked in differentsectors of the economy, while Rebelo (2005) documents that this comovement persists also
whena more disagg regatedclassification of industries is considered. My paper relates to this literature as it investigates
comovement within and between the twospecific sectors of the economy. However, the focus of this paper is on un-
derstanding the effects and relevanceof a specific shock, i.e. news shock, by studying the comovement that it induces.
©2018 The Department of Economics, University of Oxford and JohnWiley & Sons Ltd
Sectoral effects of news shocks 217
are more pronounced in durable, rather than non-durable, goods industries. Early research
in the real business-cycle tradition (see Eichenbaum, 1984; Blinder, 1986; Christiano and
Eichenbaum, 1987; Ramey, 1989) focused considerable attention on the importance of
explaining the behaviour of inventories.6In my analysis, I re-establish the role of the im-
portance of inventories with newempirical evidence concerning the response of inventories
to news shocks, connecting the two literatures.To do so, I use the inventories-to-sales ratio,
a standard inventories indicator.The resulting percentage response of inventories to news
is statistically significant in the durable goods sector, reinforcing the notion of the impor-
tance of inventories specifically in the durablemanuf acturing subsector (e.g. Feldstein and
Auerbach, 1976; Blinder and Holtz-Eakin, 1984).
This large and significant response of inventories in the durable goods sector to a
news shock suggests that the behaviour of inventories might carry relevant information for
understanding the propagation of news shocks and business cycles.7Therefore, to explore
the mechanism, I build a model with an explicit role for inventories in the durable goods
sector. Specifically, my model is a two-sector, two-factor, real business cycle model that
follows Baxter (1996) in its basic structure. Sector 1 produces a pure consumption (non-
durable) good, whereas sector 2 produces a consumer durable goods and the capital good
that is used as an input in the production of both consumption goods. Both sectors use
capital and labour as their factor inputs. The key difference between the two sectors is that
a good produced in sector 1 is perishable, whereas a good produced in sector 2 can be
stocked. I model this feature by adding inventoriesinto the production function of sector 2,
following Christiano (1988) and Kydland and Prescott (1982). These authors argue that
the stock of inventories, as the stock of fixed capital, provides a flow of services to a firm.
My model features several additional components. First, it requires adjustment costs
both in investment and in new purchases of durablegoods, because this gives the agents an
incentive to respond to positive news immediately. Second, variable capital utilization in
both sectors serves an important function by creating a channel through which hours and
output can respond to news shocks. Third, my model introduces preferences with a weak
short-run wealth effect on the labour supply.This feature plays an important role in securing
my results because the empirical evidence does not easily square with a two-sector real
business cycle model with standard preferences of the King, Plosser and Rebelo (1988)
type. While these preferences are desirable for obtaining a negative response of labour
supply at the aggregate level, they cannot generate comovement between hours worked
across the twosectors. Therefore, a model with the capacity to reproduce empirical evidence
must feature preferences with a weak short-run wealth effect on the labour supply. Finally,
as mentioned above, durable goods sector requires inventories for production.
These components together explain the observed empirical responses. Specifically, the
model can replicate the negative impact responses of hours in both sectors. The presence
6More recently, many papers have looked at the aggregate implications of introducing inventories into dynamic
general equilibrium models (e.g. Bils and Kahn, 2000; Fisher and Hornstein, 2000; Kahn, 2008a,b; Kryvtsov and
Midrigan, 2013).
7To the best of myknowledge, Crouzet and Oh (2016) is the only paper that investigates inventories dynamics in
the context of the news literature; these authors document that the dynamics of the inventories-to-salesratio is crucial
for the identification of aggregate news shocks. However,they do not investigate sectoral components of aggregate
news shock and the role of inventories in explaining these components.
©2018 The Department of Economics, University of Oxford and JohnWiley & Sons Ltd
To continue reading
Request your trial