Regime Dependent Effects of Inflation Uncertainty on Real Growth: A Markov Switching Approach

AuthorMustafa Caglayan,Ozge Kandemir Kocaaslan,Kostas Mouratidis
Published date01 May 2016
Date01 May 2016
DOIhttp://doi.org/10.1111/sjpe.12087
REGIME DEPENDENT EFFECTS OF
INFLATION UNCERTAINTY ON REAL
GROWTH: A MARKOV SWITCHING
APPROACH
Mustafa Caglayan*, Ozge Kandemir Kocaaslan** and Kostas Mouratidis***
ABSTRACT
We empirically investigate the effects of inflation uncertainty on output growth
for the United States between 1960 and 2012. Modeling output dynamics within
a Markov regime switching framework, we provide evidence that inflation uncer-
tainty exerts a negative and regime-dependent impact on output growth. A bat-
tery of sensitivity checks confirm our findings.
II
NTRODUCTION
Economists agree that sustainable growth, and low and stable inflation consti-
tute two of the fundamental objectives of the policymakers. A reason behind
this conviction is that high and unstable inflation leads to an increase in infla-
tion uncertainty distorting the efficient allocation of resources. To that end
Friedman (1977) emphasizes that (1) an increase in inflation raises inflation
uncertainty;
1
and that (2) high uncertainty, distorting the information content
of prices, hinders the efficient allocation of resources. Along these lines, Beau-
dry et al. (2001) argue that during periods of high inflation volatility manag-
ers would be unable to detect profitable investment opportunities as it is
harder to extract information about the relative prices of goods. Furthermore,
during periods of high uncertainty, external funds become prohibitively expen-
sive due to heightened asymmetric information problems causing managers to
delay or cancel fixed investment projects. Lower investment, in turn, impedes
output growth.
More recently, using structural models, several researchers have begun to
examine the channels through which uncertainty could affect real variables.
For instance, Bloom (2009) shows that macro uncertainty shocks cause a
rapid drop and rebound in aggregate output and employment as firms
*Heriot-Watt University Edinburgh
**Hacettepe University
***University of Sheffield
1
A vast empirical literature provides support for this hypothesis. See for instance Caglayan
et al. (2008) and the references therein.
Scottish Journal of Political Economy, DOI: 10.1111/sjpe.12087, Vol. 63, No. 2, May 2016
©2015 Scottish Economic Society.
135
temporarily pause their investment and hiring. Fern
andez-Villaverde et al.
(2011) show that fiscal volatility shocks reduce output, consumption, invest-
ment, and hours worked drop on impact and stay low for several quarters.
2
Basu and Bundick (2012), using a non-competitive one-sector model with
countercyclical markups, show that in response to an uncertainty shock out-
put, consumption, investment, and hours worked falls. Nakata (2012), using a
standard New Keynesian model, finds that an increase in the variance of
shocks to the discount factor process reduces consumption, inflation, and out-
put. Mumtaz and Theodoridis (2015) find that supply side uncertainty shocks
lead to lower output due to precautionary savings. Yet, other researchers, for
instance Bachmann and Bayer (2009), point out that risk might not be impor-
tant to generate business cycle fluctuations.
While it is important to examine the uncertainty effects on real variables
within the context of structural models, it is equally important to recognize
that ignorance of the underlying nonlinearities in the data will lead to biased
conclusions. Especially, if the relationship between explanatory variables and
the independent variable were to change as the state of the economy varies,
linear models would yield biased coefficient estimates and standard errors.
3
Given this particular shortcoming, some researchers have recently begun to
implement stochastic volatility models within the context of structural models
to examine the impact of uncertainty on real variables.
4
This approach,
although attractive, as Fern
andez-Villaverde and Rubio-Ram
ırez (2013) point
out, in cases where the underlying process has discrete jumps, SV model will
anticipate the changes by showing changes in volatility before they happen.
This result is due to the fact that estimation methods favor small rather than
large changes in the data. Hence, in cases where data present regime shifts, it
is advisable to use other approaches which are designed to capture such
changes in the data.
In this study, recognizing the presence of regime shifts in inflation and out-
put growth series, we examine the effects of inflation on output growth by
implementing a Markov regime switching approach. To pursue our examina-
tion, we follow a two-step approach. In the first stage, we implement a Mar-
kov regime switching GARCH model to obtain a proxy for inflation
uncertainty. In the second stage, we examine the level and the volatility effects
of inflation on output growth using a Markov regime switching framework.
One important advantage of this approach is that the model determines the
regime switches endogenously. In our investigation, we scrutinize the growth
rates of both monthly industrial production and quarterly gross domestic
product data for the United States. Our findings based on both industrial pro-
duction and GDP growth rates provide evidence that the impact of inflation
2
Also see Primiceri (2005) who discusses the persistence of uncertainty regimes.
3
Evans and Wachtel (1993) infer that models which do not account for regime changes in
the inflation process underestimate not only the extent of uncertainty but also the uncertainty
effects on economic growth.
4
See for instance Fern
andez-Villaverde and Rubio-Ram
ırez (2013) and the references
therein.
136 M. CAGLAYAN, O. K. KOCAASLAN AND K. MOURATIDIS
Scottish Journal of Political Economy
©2015 Scottish Economic Society

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