Episodes of War and Peace in An Estimated Open Economy Model

Version of Record: https://www.sciencedirect.
com/science/article/pii/S0165188919301058
Manuscript_972d1ec835c48481cae5b5af0c0aa8cd
Episodes of War and Peace in an

Estimated Open Economy Model∗
Stéphane Auray† Aurélien Eyquem‡
June 13, 2019
Abstract
This paper proposes a straight model-based analysis of historical war episodes. We analyze
the effects of world wars on the macroeconomic dynamics in the US, France, Germany, and
the UK, by means of an estimated open-economy model where war episodes are modeled as
an additional source of observed shocks. The model allows war episodes to affect the economy
through lower capital depreciation, partial default on public debt, a military draft, changes in
household preferences, and spillovers on other shocks (productivity, investment, trade, policy
variables). In the US, the bulk of fluctuations during war episodes can be mainly explained by
the rise in government spending, and the war shock plays a minor role. In other countries, the
war shock is an essential driver of fluctuations. We also discuss the size and state-dependence
of public spending multipliers, and produce a counterfactual exercise to quantify the welfare
losses from war episodes.
Keywords: Fluctuations, War, Trade, Taxes, Public Debt, Bayesian estimations, Multipli-
ers, Welfare.
JEL Classification: E31, E32, E62, F41, H56, H63.
∗
We thank the Editor and three referees for insightful comments that lead to an improved paper, as well as
participants at various seminars and conferences for helpful comments. The authors acknowledge the nancial
support of Projets Generique ANR 2015, Grant Number ANR-15-CE33-0001-01.
†
CREST-Ensai and Université du Littoral Côte d’Opale. ENSAI, Campus de Ker-Lann, Rue Blaise Pascal, BP
37203, 35172 BRUZ Cedex, France. [email protected].
‡
Univ Lyon, Université Lumière Lyon 2, GATE UMR 5824, F-69130 Ecully, France, and Institut Universitaire
de France. [email protected].
© 2019. This manuscript version is made available under the Elsevier user license
https://www.elsevier.com/open-access/userlicense/1.0/
1 Introduction
How can a dynamic general equilibrium model account for the macroeconomic effects of war
episodes like world wars? This paper proposes an approach based on the estimation of a medium-
scale model on a large set of historical macroeconomic time series for France, Germany, the UK
and the US dating back to the 19th century. The model features usual characteristics of DSGE
models, and is augmented with a variety of assumptions that help the model to account for the
macroeconomic effects of war episodes. The paper can be thought of as an ambitious extension
of papers like McGrattan and Ohanian (2010), or Devereux and Smith (2007), as it makes use
of a general equilibrium model to account for historical episodes for various countries. The chief
difference between our paper and most papers that study historical episodes is that we estimate
our model on a long historical dataset that comprises both World War I (WWI) and World War
II (WWII) episodes, the hyperinflation period of the 20’s in Germany, the Great Depression of
the 30’s, and the post-war recovery.
The model is a significant open-economy extension of the Smets and Wouters (2005) model.
It allows for trade in some components of demand (private consumption and investment), and
consider incomplete international financial markets, to account for potential wealth transfers
induced by war episodes. It also includes a more extensive set of policy instruments: the money
growth rate, public spending, public investment, public debt and distorsionary taxes on labor
and capital. We consider structural shocks that affect productivity, the efficiency of investment,
labor and capital income tax rates, public consumption, the money growth rate, and foreign
demand.1
War episodes are modeled as an additional source of shocks, that leads some parameters to switch
to (estimated) specific values and trigger some (estimated) specific effects on other structural
shocks. More precisely, we introduce an (observed) indicator variable ∆t that takes two values,
∆t = 1 during world war episodes, and ∆t = 0 otherwise. When ∆t = 1, the depreciation
rate of capital rises to capture war-related capital destruction, the possibility of a military draft
is introduced, a partial default on public debt may occur, and households’ preferences towards
the public good can be affected. In addition, when ∆t = 1, a war-specific effect on exogenous
variables such as productivity or foreign demand is considered, which provides an additional
degree of flexibility to the model. Those may capture all the features that are not explicitly taken
into account by our model, but might affect aggregate productivity, investment, trade, policy
instruments during war episodes.2 Canova, Ferroni and Matthes (2019) show that allowing for
1
We also introduce a measurement error shock on the measurement equation of GDP growth.
2
The model does not take explicitly into account all the dimensions along which war episodes affect the economy:
rationing, financial markets disruptions, changes in national borders, changes in the monetary standard – from
fixed exchange rates before WWI, to suspension during WWI, to resumption afterward, to floating rates before
WWII, to capital controls during WWII, Bretton Woods after WWII and its suspension after 1973 –, diseases,
population displacement, to name just a few.
1
time-varying nature of some parameters in DGSEs can affect substantially the decision rules and
resulting IRFs, and that time-varying or state-dependent parameters can improve the quality
of the fit of estimated models. From a methodological point of view, our approach is not a
Markov-Switching Bayesian estimation because the state (the war indicator variable) is perfectly
observed, but the idea is to capture some of the non-linearity of the data generating process due
to war episodes.
We have four different datasets for France (1898-2006), Germany (1880-2008), UK (1870-2005)
and the US (1871-2010). For each country, the dataset features a indicator variable ∆t for world
war episodes and a set of macroeconomic quantities or prices that mixes data mostly from Barro
and Ursúa (2017) and from Piketty and Zucman (2014). Models are estimated using Bayesian
methods. For each country, we obtain a set of point estimates for parameters and a set of
smoothed structural shocks that fit the observed time series. Our main goal is to understand
how the model accounts for the dynamics of the economy during war episodes. In particular, we
are interested in the estimated relative contribution of war-specific channels compared to other
sources of fluctuations. Along with McGrattan and Ohanian (2010), our historical decomposition
finds that the contribution of war-specific channels is relatively minor for the US, and that the
bulk of fluctuations during war episodes is mostly explained by large positive public spending
shocks. For France and Germany, the contribution of shocks on the war indicator variable is
much more important relative to other types of shocks. Historical decompositions for the UK are
somewhere in between those two situations: shocks on the war indicator variable are important
in explaining the investment ratio and net exports, but not so much GDP growth rate or the
inflation rate. Hence, the “macroeconomic of war and peace” can not solely rely on government
spending shocks and productivity shocks for countries other than the US. Capital destruction,
negative productivity and investment spillovers from war episodes, foreign demand dynamics,
war-specific tax dynamics all matter substantially, at least for countries like France or Germany.
Based on our estimations, we also investigate the effects of simulated war episodes on the dy-
namics of the four countries. In the US, we find that war episodes are mostly driven by the
positive public spending spillover. Other channels (draft, capital depreciation, default) and
other spillovers play only minor roles. To some extent, the dynamics for the UK is similar to
that of the US since the government spending spillover plays a key role too. For the UK how-
ever, other important features have significant effects: war episodes reduce productivity, hurt
investment efficiency, and produce a disruption in external trade, captured by a fall in foreign
demand. France and Germany feature different responses. In these countries, our estimations
find the public spending spillover to be statistically non-significant but productivity and invest-
2
ment spillovers are negative and significant.3 As a consequence, a simulated war episode lowers
GDP by 40%, and consumption falls more than GDP. Channels like the depreciation rate of
capital also matter more than in the US or in the UK. For France, the trade spillover is also
negative and very large.
Our main conclusions thus complement Braun and McGrattan (1993) and McGrattan and Oha-
nian (2010). According to them, a neoclassical model fed with large public spending shocks
explains most of the dynamics of the economy during world war episodes in the US or in the
UK.4 We find similar results for the US and, to some extent, for the UK. However, for France
and Germany, the macroeconomic effects of war episodes can not be reduced to the sole massive
rise in public spending. We find that productivity, investment or trade spillovers, and the shock
on the war indicator variable in general, are important for those countries.
Our paper echoes a series of work suggesting that these are important aspects of macroeconomic
dynamics during war episodes. In particular, Auray, Eyquem and Jouneau-Sion (2014) show
that capital depreciation shocks à la Ambler and Paquet (1994) crucially contribute to the
macroeconomic dynamics induced by major war episodes.5 Another part of the literature also
highlights the importance of external trade and finance to account for the dynamics of economies
during wars.6 More generally, war episodes most often induce large disruptions in trade, that
contribute to the dynamics of key domestic macroeconomic variables (see Anderton and Carter
(2001) or more recently Glick and Taylor (2010)). These channels are potentially captured by
our flexible model assumptions.7
Using our estimation results, we also investigate the question of state-dependent public spending
multipliers. Indeed, many papers use war episodes to identify the effects of discretionary public
spending shocks, the “narrative approach”.8 It is therefore crucial to know whether the resulting
multipliers are the same during war episodes and during normal times. We find that the impact
and short-term output multipliers are not significantly different in the US and in the Germany.
For France and the UK, we find that impact and short-run output spending multipliers are
3
Of course public spending rises in the data in these countries, but the estimations capture this rise by means of
positive innovations to public spending rather than by estimating a positive spillover from war episodes. This may
come from large differences in the spending patterns during WWI and WWII, or from differences in the timing of
spending increases.
4
An additional contribution is Ohanian (1997), who shows the great importance of financing schemes in the
effects of war episodes on output and welfare.
5
See also Furlanetto and Seneca (2014) regarding the importance of capital depreciation shocks over the business
cycle in normal times.
6
For example, Devereux and Smith (2007) analyze the 1871 Franco-Prussian war indemnity through the lens
of the transfer problem, invoking the 1929 Keynes-Ohlin controversy.
7
Focusing on a different issue, Martin, Mayer and Thoenig (2008) identify the impact of international trade on
the occurrence of conflicts. To sum up, data show that the sign and the magnitude of the relationship between
trade openness and armed conflicts depend on the specific characteristics of trade flows and agreements but that
major conflicts reduce international trade flows.
8
See Ramey and Shapiro (1997), Ramey (2011b), Ramey (2011a), and more recently Ben Zeev and Pappa
(2015) and Ramey (2019).
3
significantly larger during war episodes than during normal times, but that long-run multipliers
tend to converge. Our results point to the possibility that output spending multipliers differ
during war episodes, at least for some countries. Finally, we also quantify the welfare losses from
war episodes and contrast those to the welfare losses from fluctuations. Unsurprisingly, we find
that war episodes generate large welfare losses relative to other shocks, especially for Germany
and France.
The paper is organized as follows. Section 2 describes and justifies our original model assump-
tions. Section 3 presents our datasets. Section 4 presents the estimation strategy. Section 5
reports our estimates for the US, discusses the resulting historical decomposition and analyzes
the state-dependence of public spending multipliers. Section 6 compares our results for the UK,
France and Germany to those derived for the US. Section 7 presents an analysis of the welfare
losses from business cycles, as well as a quantification of the specific welfare losses generated by
war episodes. Section 8 concludes.
2 Datasets
We have four separate datasets respectively for France (1898-2006), Germany (1880-2008), UK
(1870-2005) and the US (1871-2010). For each country, a first time series is an indicator variable
for war periods ∆t that takes two values: ∆t = 1 during world war episodes, and ∆t = 0
otherwise. In the model, this variable will be treated as an exogenous shock.
In addition to the war indicator variable, for each country, the dataset includes GDP and con-
sumption per capita in 2006$, investment as a percentage of GDP (it /yt ), total consumption
expenditure (public and private) as a percentage of GDP ((ct + gt ) /yt ), net exports as a per-
centage of GDP (nxt /yt ), the public debt-to-GDP ratio (brt /yt ), total tax receipts as a percentage
of GDP and the PPI or CPI inflation rate (πd,t or πt ). GDP and consumption per capita are
taken from the dataset of Barro and Ursúa (2017). Other variables are taken mostly from Piketty
and Zucman (2014). For the US, total tax receipts are taken from Mitchell (1998) before 1947
and the FRED database after 1947. For Germany, total tax receipts are taken from Flandreau
and Zumer (2004) before 1913 and from Piketty and Zucman (2014) from 1950 onwards. The
inflation rate is based on the CPI for France and Germany, and on the GDP deflator for the UK
and the US, a distinction that our open-economy model handles easily.9
Our datasets are made stationary in the following way. Some variables are considered in growth
rates and demeaned. We thus consider the demeaned log-difference of GDP (dytobs = ∆log(yt ) −
∆log(yt )), consumption (dcobs
t = ∆log(ct ) − ∆log(ct )) and inflation (πtobs = ∆log(pt ) − ∆log(pt )
for France and Germany, and πtobs = ∆log(pd,t ) − ∆log(pd,t ) for the US and the UK). We take
9
Over the whole sample, it was not possible to obtain a GDP deflator for the four countries.
4
public debt stocks and tax receipts in levels combining the ratios with the levels of GDP, and
take the demeaned log-differences (dbobs
t = ∆log(brt )−∆log(brt ) and dtobs
t = ∆log(Tt )−∆log(Tt )).
Other variables, total consumption expenditure (cgytobs ), net exports (nxytobs ) or investment
(iytobs ) are expressed relative to GDP. Over the whole sample, those variables display trends,
mostly because the share of public spending in GDP has risen substantially over time, lowering
the share of private investment and consumption in GDP. We make these ratios stationary
by removing a linear-quadratic trend. Whenever our datasets have missing points, they are
handled as follows.10 Series are first interpolated with a spline before being transformed (taken
in log-difference or having their trend removed). Then the interpolated missing observations
are removed, to be properly handled as missing observations by the estimation algorithm (see
Section 4). Appendix A reports the time series used to estimate the model.
For France and the UK, the reported time series show that war episodes significantly depressed
the economy: the GDP growth rate and the consumption growth rate fell to historical lows,
consumption growth fell as much as or more than the GDP growth, inflation was very high,
the investment-to-GDP ratios reached historically low levels, the total consumption expenditure
to GDP ratios were the highest in history, because public expenditure exploded. Net exports
dropped dramatically and public debt jumped to very high levels. For these two countries, war
periods represent the largest swing in the reported time series, by far.
During WWI, France and the UK went to war simultaneously in the summer of 1914. France and
the UK faced Germany on the front in eastern France and in the seas, in northern France and in
the North Sea. The UK decided to impose a maritime economic blockade on Germany. From 1915
to 1917, the conflict was mostly a ”Position war”, where belligerents were completely immobile
for months. In July 1918 occurred the major ”Second Marne Battle”, that saw the Allies and
France in particular inflict great losses to Germany, thereby ending the war in November 1918.
This quick summary suggests that France and the UK had relatively similar war experiences
during WWI.
As for WWII, it was officially declared when Germany invaded Poland in 1939, an ally of France
and the UK. However, they did not react during the first months. In May and June 1940, France
was invaded by Germany, but the UK fought Germany in the air and seas. From September
1940 to May 1941, the UK suffered many bombings, leading to large destruction in London and
Coventry. Germany and the UK also fought in the Mediterranean Sea, mobilizing a lot of UK
military resources. After the US entered the war in 1942, they supplied the UK with military
and civilian equipments and current consumption goods. The Allies finally broke the Atlantic
10
For Germany, investment and consumption expenditure to GDP have missing points from 1914 to 1924, and
from 1939 to 1949. Net exports to GDP have missing points from 1919 to 1924 and from 1939 to 1949. Inflation
data in 1922 and 1923 are considered missing. Total tax receipts are missing from 1914 to 1949. For the US, tax
receipts are missing from 1918 to 1924.
5
Wall in June 1944, during the Normandie Landings, and freed France (among others). France
and the UK therefore had quite different involvements in WWII, France was at war only for a few
months, lost almost the entirety of its military equipments to Germany, and was occupied until
late 1944. The UK was at war almost continuously from 1939 to 1945, and obtained substantial
help from the Allies (the US in particular) starting in 1942.
The dataset for Germany shares some of the features highlighted for France and the UK. During
world war episodes, GDP and consumption growth fell substantially, net exports were highly
depressed (during WWI). Due to the abundance of missing data points, the picture remains
difficult to square completely with the narratives for France and the UK. In addition, the post-
WWI reparation payments and the associated subsequent hyperinflation episode in the early
20’s is an additional major swing in the time series of Germany, in particular when looking
at skyrocketing inflation and plunging public debt growth. Further, the Great Depression of
the early 30’s also plays a major role in the observed macroeconomic time series: this episode
witnessed the lowest investment-to-GDP ratio and the highest total consumption to GDP ratio
in the history of Germany, although it must be noted that data points are missing during war
episodes for both variables.
During WWI, Germany had to fight the Soviet Union, France and the UK. They suffered an
economic blockade after the UK controlled the North Sea, and eventually lost WWI. Histori-
ans believe that WWII is widely rooted in the severity of the Versailles Treaty that resulted
from WWI, and in the 1929 crisis that led many countries to apply protectionist measures and,
for Germany in particular, rearmament policies to restore economic growth.11 During WWII,
Germany invaded many countries and seized a large quantity of military and civilian resources.
However, given that Germany was on three fronts simultaneously, the economy had a hard time
producing enough. In addition, the building of the Atlantic Wall was expensive in terms of labor
and construction material. Late 1944 and 1945, Germany was severely destroyed by bombings,
and invaded by the Soviet Union and Allied Forces. The amount of destruction (housing, facil-
ities, infrastructure) undergone in 1945 was massive, as suggested by the large observed fall in
GDP growth. During both world wars, Germany was fighting all along, on home and foreign
soils, experienced and inflicted great amounts of human casualties and material destruction.
The dataset for the US shows a different pattern, as war episodes depressed consumption growth
but saw a rise in GDP growth. The investment ratio fell during world wars while the total
expenditure ratio rose. However, historical lows/highs for these variables are reached during
the Great Depression of the 30’s. Similarly, the lowest GDP and consumption growth rates are
observed during the Great Depression. The pattern of net exports is also quite different: net
exports went up dramatically during WWI and after 1942 during WWII. Public debt growth
11
To a certain extent, the involvement of Germany in the 1936 Spanish Civil War, the Anschluss and the Sudeten
conflict in 1938 all suggest that Germany was already at war prior 1939.
6
jumped quite significantly during both world wars, in line with the fact that public expenditure
rose to historical highs during both war episodes. Notice also that inflation rose to high levels
during war episodes, especially during WWI. One of the reasons the US had different economic
dynamics during world wars is most certainly due to the late involvement in both world wars,
and the relative preservation of the US soil with respect to combats and destruction.
Indeed, the US remained neutral for a long time before entering WWI in June 1917. They offered
a massive support in terms of military divisions (representing up to 31% of Allied human forces),
as well as marine forces. Their support was key to win the war, and in particular the Battle of
the Atlantic, during which Germany tried to destroy commercial ships headed to the UK, as a
retaliation against the economic blockade. The US also entered WWII quite late, in 1942, and
initially focused their military resources on the Pacific front. Later in 1942 and in 1943, the
US started supplying the UK in goods and equipments, and ended-up engaging great amounts
of resources until 1945 to help win the war. Overall, the US experienced little destruction on
US soil during both wars, but consumed a lot of resources in terms of military equipments, and
providing numerous armed forces through massive military drafts.
3 The model
We consider a small open economy model with a government in charge of fiscal and monetary
policy. The model is based on a standard DSGE such as Smets and Wouters (2003) with sticky
prices and wages, but is extended along many dimensions. First, it considers an open economy
with trade in consumption and investment goods, home bias, and an incomplete international
financial market. These assumptions are introduced to account for trade balance dynamics and
potential wealth transfers. Second, it introduces money in the utility function and monetary pol-
icy is not conducted through a Taylor-type rule, but is affected by exogenous money growth rate
shocks. In addition, seigniorage revenues are transferred to the government.12 Both assumptions
are consistent with a non-trivial role for monetary policy and for a potential financing of public
spending through seigniorage revenues, a feature that is likely to happen during or after war
episodes. Third, the government levies distortionary taxes on labor and capital income. Both
tax rates are assumed to follow exogenous processes reflecting discretionary decisions made by
the government. The government spends on domestic consumption goods, invests in domestic
public capital goods, issues government bonds, levies taxes and has access to seigniorage revenues
by issuing money. These features allow us to consider a set of shocks that is quite different from
12
Several contributions, starting with Friedman and Schwartz (1963), highlight the consistency of exogenous
money growth rules with monetary facts (see also Monnet and Weber (2001)). Further, observational equivalence
under a variety of assumptions has been extensively proven in the literature on monetary policy. Finally, a
substantial amount of Central Banks’ actions in the early 20th century can be understood by examining their
international role in defending the Gold Standard (see Fishback (2010)). This is particularly the case for the Fed
during the 30’s, and motivates our modeling of monetary policy decisions as exogenous innovations to the money
growth rate.
7
those considered by Smets and Wouters (2003), as we consider tax shocks and leave preference
or mark-up shocks aside. Finally, we introduce an indicator variable that is ∆t = 1 during war
episodes and ∆t = 0 otherwise. When ∆t = 1, the depreciation rate of public and private capital
stocks is larger, the possibility of a military draft is opened, a partial default on public debt can
arise, and the elasticity of substitution between public and private goods in households’ pref-
erences can take a different value. The next paragraphs discuss the key departures from usual
open-economy DSGE models while the details of the model are given in Appendix B.
We consider a unit continuum of households indexed in j that maximize their lifetime welfare
(∞ !)
X
s−t cs (j)1−σc (ms (j) /ps )1−σm ns (j)1+ψ
Wt = Et + χm − χn (1)
e
β
s=t
1 − σc 1 − σm 1+ψ
In the welfare index, total consumption e

cs (j) is a bundle of individual consumption of the private
good ct (j) and the total consumption of public good gt
νt −1
νt −1 νt −1 ν
t
νt
ct (j) = ϕct (j) t + (1 − ϕ) gt
e ν , νt > 0 (2)
When νt > 1, private and public goods are substitutes while νt < 1 implies complementar-
ity.13 We introduce the possibility that this parameter takes different values whether countries
experience a war episode or not, that is
νt = (1 − ∆t ) νnorm + ∆t νwar (3)
where ∆t = {0, 1} is an indicator variable that equals one during war episodes and zero in normal
times. Endogenous labor supply and money in the utility function are introduced to account
respectively for labor market and money market dynamics over the business cycle in normal
times and during war episodes.
The budget constraint of agent j is relatively standard. She has access to four assets: money,
private capital, public debt and foreign assets. The originality of our model is to allow for a
war-specific partial default on public debt, and for a war-specific depreciation rate of private and
public capital. Formally, these war-specific features are triggered by the war indicator variable
∆t . The first assumption intends to capture the potential malfunctioning of public bond markets
during war episodes, or some kind of fiscal repression. According to the second one, capital
depreciation is higher during war episodes:

δt = δ 1 + ∆t pδ , pδ > 0 (4)
13
See Bouakez and Rebei (2007) and references therein for a discussion of the empirical relevance of this as-
sumption.
8
As in Auray et al. (2014), this specification captures some of the capital destruction through
higher capital depreciation, where pδ governs the size of war-related destruction. Notice that the
time-varying depreciation rate of capital applies to both private and public capital.
In addition, capital accumulation is subject to investment adjustment costs and to investment

shocks ζt . These shocks evolve according to an autoregressive process, augmented with a specific
effect during war episodes. As in Justiniano, Primiceri and Tambalotti (2011), they may capture
all the potential sources of frictions that affect the transformation of savings into future capital
input.
ζt = (1 − ρζ ) + ρζ ζt−1 + φζ ∆t + εζt (5)
The idea of the war-specific spillover φζ is to account for the potential extra kick in private capital
accumulation that war episodes might generate through expropriation, changes in the purpose
of production facilities, disruption in aggregate private demand, banking or financial frictions,
or rising uncertainty in general.
The derivation of the first-order conditions is standard and thus reported in Appendix B. It is
useful however to point that we allow for trade in consumption and capital goods, as those goods
are made of domestic (d) and foreign (f ) goods. The bundle has a CES specification allowing for
home bias and a non-unitary elasticity of substitution. Further, our open-economy framework
allows for incomplete international asset markets, to account for potential wealth transfers that
might occur during war episodes, or even during normal times.
There is a continuum of final good producers indexed in ω, with the following production function
1−ι
yt (ω) = at kts (ω)ι 1 − ∆t p` `t (ω) , p` > 0 (6)
where kts is a measure of capital services used in production — we allow for a variable utilization
rate in the optimization program of households — and 1 − ∆t p` `t is the amount of labor that

firms use to produce. Total labor `t is affected by a potential military draft that lowers the amount
of labor used by private firms during war episodes by p` . Finally, at is a productivity measure
following an autoregressive process. Again, we introduce the possibility that war episodes have
extra effects on productivity through channels that are not explicitly specified in the model
through a war spillover φa :
at = (1 − ρa ) + ρa at−1 + φa ∆t + εat (7)
Those factors might relate to shift in the type of products that firms produce (from regular
industrial goods to weapons or military equipments), capital and labor reallocation among firms,
changes in the composition of the labor force, congestion in transport/public infrastructure, bad
communication between headquarters and other establishments of the same firm, among other
9
things.
The government is in charge of both monetary policy, public spending and fiscal policy. Its budget
constraint is

bt + mt + Tt = rt−1 1 − ∆t pb bt−1 + mt−1 + ∆t p` wt `t + pd,t (gt + igt ) (8)
where
Tt = τtn wt nt + τtk rtk zt − pt δt kt−1 + taxt (9)
stands for total tax receipts. The left-hand side of Equation (8) shows government revenues and
the right-hand side features government expenditure. In particular, the government has to pay
drafted people.14 Following Leeper (1991), we model fiscal policy as a set of feedback rules for
taxes and government spending, which ensures that the intertemporal government budget con-
straint holds and that the equilibrium is determinate. These rules imply that policy instruments
react to the lagged deviations of real public debt brt to its steady-state value. In addition, we con-
sider war-specific spillovers on taxes and spending through φg , φτ n and φτ k . Those are intended
to capture war-specific decisions made by governments in terms of public spending, tax policies,
and more broadly war-specific government financing schemes. Further, policy instruments are
also affected by autoregressive exogenous shocks:
gt = g + dτn (brt−1 − br) + φg ∆t + sgt (10)

τn
τtn = τ n + dτ n (brt−1 − br) + φτ n ∆t + st (11)
τk
τtk = τ k + dτ k (brt−1 − br) + φτ k ∆t + st (12)
where
sgt = ρg sgt−1 + εgt (13)

sτt n = ρτn sτt−1
n
+ ετt n (14)
sτt k = ρτk sτt−1
k
+ ετt k (15)
We consider that money supply grows at a rate µm

t = mt /mt−1 that also follows an autoregressive
process augmented with a war-specific spillover φµ :
µm m m
t = (1 − ρm ) + ρm µt−1 + φµ ∆t + εt (16)
The latter is introduced to account for potential war-specific seigniorage to monetize public debt.
Further, as in Auray et al. (2014), the government adjusts the amount of public investment to
14
Here we follow the assumption made by Siu (2008) in its baseline model according to which wages paid to the
military equal market wages.
10
secure a given and constant level of public capital k g :
igt = δt k g (17)
which will trigger a rise in public investment during war episodes, capturing the need to rebuild
destroyed military equipment, facilities, buildings, or public infrastructure (roads, bridges, etc...).
A final source of fluctuations in the model is foreign demand. The latter enters the goods market
clearing condition and raises or lowers the demand for domestic goods that are exported to the
rest of the world (See Equation (B.40) in Appendix B). Foreign demand is considered exogenous
to the economic conditions of domestic economy, and is affected by an autoregressive shock,
augmented with a war-specific component. The latter intends to capture the potential disruptions
in trade and/or wealth transfers the war episode might produce through global demand slowdown,
a restriction of the use of transportation infrastructure for the purpose of private trade, among
other things. The war spillover could as well boost trade, for instance if the country is invaded
and the occupant massively shifts local resources to its own country, or if the occupied country
produces goods or equipment for the occupant.
Overall, the model features the following structural shocks: the productivity shock εat , the in-
vestment shock εζt , the foreign demand shock εyt ∗ , the tax shocks ετt n and ετt k , the public spending
shock εgt , the money growth shock εµt and the shock on the war indicator variable ∆t . The
latter is particular in that it introduces a war-specific regime in which capital depreciates more
quickly by pδ , public debt is partly defaulted on by a fraction 1 − pb , a fraction p` of the labor

force is being drafted, and the degree of substitutability between the private good and the public
good νt in the utility of households is affected. Finally, the shock on the war indicator variable
potentially triggers spillovers on all the other driving forces of the model.
We believe that the introduction of the war indicator variable is able to capture some of the
disruptions in economic activity that occur during war episodes: destruction of human lives,
private and public facilities and military equipment, changes in consumption dynamics through
public restrictions on consumption goods, higher public spending in consumption or capital
goods, lower aggregate productivity, brutal changes in tax rates or tax receipts, seigniorage,
partial default, large trade/global demand disruptions, wealth transfers through expropriation
and spoliation of national goods when a country is invaded, etc... In addition, note that the
effects of war episodes, triggered by a shock on ∆t , on other shocks through spillover terms can
be persistent, thereby affecting the dynamics of exogenous variables long after war episodes are
over.
Nevertheless, the war indicator variable and the spillover terms do not necessarily accurately
account for all the effects of world war episodes, and other shocks might be important. For
11
instance, if seigniorage was used after WWII but not after WWI in a given country, the spillover
coefficient might not be significantly different from zero in the estimation, and the innovations
on money growth will play a major role. As such, the war indicator variable can not integrally
account for the economic dynamics during war periods. All shocks, including the war indicator
variable, might be of importance. We argue however that our shocks and model assumptions
are flexible ways of taking into account the complex dynamics and interactions that may arise
during war episodes as well as during normal times. Our approach here can therefore be brought
in relation with the proposal of Chari, Kehoe and McGrattan (2007), who argue that models can
be used to uncover the dynamics of wedges: exogenous shocks may capture some of the dynamics
from the data that are not modeled explicitly.
4 Estimation
We estimate our model using Bayesian methods, adopting the standard approach of An and
Schorfheide (2007). This implies obtaining the posterior distribution of our estimated parameters
based on the linear approximation of the model’s solution around the steady state using the
Kalman filter. A major advantage of the approach is that it allows for the extraction of the
dynamics of shocks, as well as the historical paths of endogenous variables. We will therefore
have a complete quantitative evaluation of the model with respect to the data. In the model
section, we introduced an indicator variable ∆t for war episodes. This variable is treated as
an exogenous shock, and the war indicator variable is included in the set of observed variables.
More precisely, for each country, let ∆obs obs = 1 during war
t denote an indicator variable that is ∆t
episodes and ∆obs
t = 0 otherwise, and impose ∆t = ∆obs
t as one of the measurement equations.
Steady state. We analyze the linearized dynamics of the model around a symmetric steady-state
without inflation, implying zero net foreign assets.15 We consider that the steady state of the
economy corresponds to normal times, and we thus impose ∆ = 0. Our goal is to identify the
sources of macroeconomic fluctuations with a specific interest in war episodes, focusing on France,
Germany, the UK and the US. The model is flexible enough to allow for cross-country differences
and remains relatively agnostic on the behavior of the government. Aside from the fiscal rules,
policy instruments are allowed to feature a war-specific reaction, with no sign restrictions imposed
on those spillovers from war episodes when estimating them.16
Data vs. model mapping. For each country, a version of the model is estimated using the country-
specific dataset described in Section 2. Given that measurement error in the data is very likely,
we add a measurement error shock on the GDP measurement equation, that is specified as an
15
Details about the steady state are given in Appendix B.
16
More precisely the prior distributions considered for spillovers are Normals with a prior mean of zero, allowing
the algorithm to visit positive and negative values when computing the mode and running MH replications.
12
autoregressive process with zero mean and iid innovations:
dy obst = log(yt ) − log(yt−1 ) + yterr (18)

yterr = ρerr yt−1
err
+ εerr
t (19)
where dy obst is the observed GDP growth rate and yt is the model-based measure of GDP. Overall,
including the measurement error shock, the model features 9 shocks. To avoid stochastic singu-
larity, an estimation of the model requires at least 9 time series. For each country, the dataset
includes 8 macroeconomic time series (see Section 2) and the observed war indicator variable,
providing a total of 9 observed variables. Let Θs denote the vector of structural parameters to
be estimated, Θc the vector of calibrated parameters, and Xt a vector of endogenous variables.
The state-space form of the model writes
Xt = A (Θs , Θc ) Xt−1 + B (Θs , Θc ) εt , where εt ∼ N (0, Σ) (20)
where
n k ∗
εt = [εat , εζt , εgt , ετt , ετt , εµt , εyt , ∆t , εerr
t ] (21)
is the vector of innovations to the 7 structural shocks, the war indicator variable and the mea-
surement error shock. The measurement equation is given by17
Mt = C (Θs , Θc ) + D (Θs , Θc ) Xt (22)
where n oT
MT = dytobs , dcobs
t , πt
obs
, dbobs
t , dtobs
t , cgy obs
t , nxyt
obs
, iyt
obs
, ∆ obs
t (23)
t=1
is the vector of observed variables. The estimation procedure consists in finding the posterior
distribution P (Θs |MT ) of the parameters conditional on the data sample MT exploiting the
DSGE model likelihood function L (MT |Θs ) and the prior distribution of parameters P (Θs )
P (Θs |MT ) ∝ L (MT |Θs ) P (Θs ) (24)
The likelihood function L (MT |Θs ) is computed using the Kalman filter using
T
Y
L (MT |Θs ) = L (Mt |Mt−1 , Θs ) (25)
t=1
where Mt is the set of observed variables at time t and Mt−1 stands for the information set from
period 1 to t − 1.
17
The set of estimated equations, including the measurement equations, is given at the end of Appendix B.
13
We use the Dynare built-in estimation routine. The latter copes with missing observations
treating them as unobserved states and using the Kalman filter to infer their values during
estimation. Our goal is to use the estimated model to fill in the gaps in time series rather than
using the interpolated data, because we believe that our model does a better job in filling these
gaps than simple splines. This is important since missing data, especially for Germany, mostly
pertain to periods of war episodes.
Calibrated parameters. Before estimating the model, we fix the values of some of the parameters,
either to preserve the model’s steady state or because of weak identification from the data.18
The vector of calibrated parameters that are common to all countries is:
Θc1 = [β, ψ, σc , σm , δ, φz , ι, α, φf , χm , χn , θw ]
The time unit is the year. The steady-state discount factor is β = 0.96, producing an average
4.17% real interest rate. The risk-aversion parameters on (total) consumption and real money
balances are respectively σc = 1.5 and σm = 1.5. The steady-state capital depreciation rate is
δ = 10%, the adjustment cost on utilization is φz = 0.15 and the capital share is ι = 0.36. The
import share is α = 0.15 and the adjustment cost on foreign assets is φf = 0.025, as in Ghironi
and Melitz (2005). The value of the labor disutility parameter χn is adjusted to normalize
hours of work to one, which makes our choice for the calibrated value of the wage mark-up
(1/ (θw − 1) = 20%) basically inessential (see Appendix B for details). The Frisch labor supply
elasticity 1/ψ is calibrated to 2/3, a rather consensual value. Finally, the real balances utility
parameter is set to χm = 0.05, to produce steady-state real balances to GDP between 0.2 and
0.25, in accordance with (unreported) evidence from our dataset. Those parameters remain fixed
across countries and their values are summarized in Table 1 below.
Table 1: Calibrated common parameter values (Θc1 )
Discount factor β = 0.96
Frisch elasticity on labor supply 1/ψ = 2/3
Risk-aversion (consumption) σc = 1.5
Risk-aversion (real money balances) σm = 1.5
Capital depreciation δ = 0.10
Utilization adjustment cost φz = 0.15
Capital share ι = 0.36
Import share α = 0.15
Adjustment cost on foreign assets φf = 0.0025
Real money utility parameter χm = 0.05
Wage mark-up 1/ (θw − 1) = 0.2
Labor disutility parameter χn adjusted to get n = 1
Country-specific calibrated parameters. Remaining calibrated parameters are country-specific, to

18
We did run some pre-sample identification tests before estimation and those revealed that parameters such as
the inverse of the Frisch elasticity were potentially weakly identified.
14
account for the observed differences in the tax systems, the levels of public consumption to GDP
(κ) or public debt to GDP. The vector of country-specific calibrated parameters is
h i
Θc2 = κ, κi , τ n , τ k , θp
Our calibrated values are consistent with the raw datasets, that contain useful information about
key ratios. We also use some of Piketty and Zucman’s data that are not used to estimate
the model, such as the average capital or labor income tax rates when those are available.
Using average values of tax rates in France, we set the steady-state tax rates to τ n = 0.3 and
τ k = 0.22. The average government investment rate, taken from French data is κi = 0.0253. We
adjust public consumption to GDP κ = 0.1346 to match exactly the average level of debt to
GDP (75%) over the sample. Last, we set θp = 6 to match average net profits (20%). In the
absence of empirical evidence for profits in Germany, we set θp = 6. Data for Germany suggest
κi = 0.0164 and κ = 0.1490. The average level of capital income taxation is τ k = 0.2. We adjust
τ n = 0.2926 to match the average level of debt to GDP (45%) over the sample. Concerning
the UK, (unreported) evidence suggest an average profit rate of 25%, implying θp = 5, and the
average government investment-to-GDP ratio is κi = 0.0263. We assume τ k = 0.2 and τ n = 0.3,
and adjust κ = 0.1138 to hit the 97% average debt-to-GDP ratio featured in the raw dataset.
Finally, for the US, based on the raw dataset, we set κi = 0.0190, impose κ = 0.10 with τ k = 0.35
and adjust τ n = 0.1678 to hit the average debt-to-GDP ratio featured in the raw dataset (55%).
Those parameter values are summarized in Table 2 below.
Table 2: Calibrated country-specific parameter values (Θc2 )

France Germany UK US
Share of public investment in GDP (κi ) 0.0253 0.0164 0.0263 0.0190
Share of public consumption in GDP (κ) 0.1346 0.1490 0.1138 0.1000
Labor income tax rate (τ n ) 0.3000 0.2926 0.3000 0.1678
k
Capital income tax rate (τ ) 0.2200 0.2000 0.2000 0.3500
Steady-state price mark-up (1/(θp − 1)) 0.2000 0.2000 0.2500 0.2000
Estimated parameters. Remaining parameters of the model are estimated. Our prior distributions
are as follows. The Calvo parameters on prices η p and on wages η w are Betas, with prior means
0.25 and standard deviations 0.1. The Edgeworth complementarity parameters (νnorm , νwar ) are
estimated separately but with the same priors. They are Inverse Gammas with prior means
0.5 and standard deviations 0.25. The war-specific partial default parameter pb , the war-specific
draft parameter p` and the increase in the depreciation rate of capital induced by war episodes pδ
are all assumed to be Betas with prior means 0.1 and standard deviations 0.05. Following Smets
and Wouters (2003), the investment adjustment cost parameter φi is a Normal with prior mean 5
and standard deviation 0.5. As in the standard international RBC literature (see Backus, Kehoe
and Kydland (1993) or more recently Auray, Eyquem and Gomme (2016)), the trade elasticity
is a Normal with prior mean 1.5 and a 0.25 standard deviation. Parameters of the government
15
spending and tax rules (dg , dτn , dτk ) have positive supports to avoid indeterminacy as much as
possible. We assume they are Inverse Gammas with prior means 0.25 with a 0.25 standard devi-
ation. We are completely agnostic about the sign of war spillovers on exogenous processes, and
assume that (φa , φg , φς , φy∗ , φτn , φτk , φµ ) are Normals with prior means 0 and standard devia-
tions 2. Finally, the persistence of forcing processes (including the measurement error shock on
GDP) are Betas with prior means 0.7 and standard deviations 0.2 and the standard deviations
of innovations are Inverse Gammas, with prior means 0.1 and infinite standard deviations.
Estimation results. Tables 5 to 8 in Appendix C summarize the prior distributions and report
the means of posterior parameter distributions along with 90% confidence intervals for each
of the four countries, based on 250 000 replications of the MH algorithm where the first 20%
were discarded and where the scale parameter was adjusted to get a 30% acceptance rate. The
estimated parameter values will be discussed in the next Section, starting with the case of the
US and then contrasting the results for other countries. The Tables report the estimates for the
baseline model with all channels through which war episodes might affect the economy, and a
for a restricted model where all these channels are shut down: war spillovers and war specific
parameter (pδ , p` , and pb ) are restricted to zero, and the Edgeworth parameter is assumed not
to switch during war episodes, νwar = νnorm . By contrast with the baseline model, the restricted
model provides information about the extent to which the war indicator variable ∆t helps fit the
data for the different countries.
5 United States: estimation results and implications

5.1 Estimation results
Prior and posterior distributions are reported in Appendix D and show that the estimation proce-
dure provides enough information to update prior distributions. Further, Table 8 in Appendix C
shows that most parameters found in other estimated models have comparable point estimates.
For example Calvo parameters for the US are 0.4 for prices and 0.2 for wages, suggesting an
adjustment frequency of prices and wages between 5 and 7 quarters, numbers that are consistent
with estimates based on quarterly datasets and models. Similar comments are in order for pa-
rameters such as the investment adjustment cost, the trade elasticity or tax and spending rule
parameters.
We rather focus on the Edgeworth complementarity parameter ν, and war-specific parameters

like spillovers on shocks, the size of war-related capital depreciation, draft or default on public
debt, because they have critical impact on the dynamics of the economy during war episodes.
The Edgeworth parameter has been shown to be crucial for the size of government spending
multipliers (see Bouakez and Rebei (2007) or more recently Leeper, Traum and Walker (2017)).
16
Our estimation for the US favors a mild complementarity between public and private goods in
households’ utility function. The estimated values νnorm and νwar are close, and roughly equal
to 0.5 – recall, substitutability is observed when ν > 1.19 For the US, we can not conclude
that these estimates are statistically different, as confidence intervals overlap. Based on post-war
Canadian data and using ML methods, Bouakez and Rebei (2007) estimate a value of 0.3320
for this parameter. Larger values found by our empirical work reflect a lower complementarity
between private and public goods in the US over our sample, resulting in lower estimates for
spending multipliers.
For the US, the estimated value of p` , the size of the war draft, is 6%. It has to be compared
with the numbers reported by Siu (2008): a 6.8% conscription in the US during WWI and a
15.6% conscription during WWII. Alternative sources point to a 11% draft during WWII. Our
estimate is therefore in the range of empirically realistic values.
The estimated value of pδ , the size of war related capital destruction, is 9%. According to Young
and Musgrave (1980) the net private capital stock fell by 2.2% between 1940 and 1945 in the US
(see their Table 1.A.1), less than our estimated number. However, the observed 2.2% number
is over a 5-years period: in our model a 9% increase in capital depreciation lowers the capital
stock by 9% on impact but produces a subsequent rise in investment. Over a 5-years period,
the capital stock should therefore not fall as much as 9%. In addition, our specification of the
capital destruction parameter also accounts for the destruction of public capital. Overall, our
point estimate is thus broadly consistent with the numbers reported by Young and Musgrave
(1980).
Concerning the partial default parameter pb , our model finds an average 12% depreciation of
public financial assets during war episodes. We believe that this feature captures the rise of
default risk or uncertainty perceived by the households.20
Let us now examine the sign and size of war-specific spillovers. Our estimation for the US shows
that neither the spillover on productivity or investment are significantly different from zero at the
90% confidence level. However, the trade spillover is different from zero and positive: a “typical”
war episode would raise foreign demand for US goods, that should result in an improvement of
the trade balance. Our estimation finds that public spending respond positively and significantly,
that the labor income tax responds positively while the capital income tax responds negatively.
19
The assumption of Edgeworth complementarity/substitutability does not necessarily capture deeply micro-
founded features of households’ preferences but we do think that it captures quite accurately the transmission
mechanism of public spending shocks to private consumption. The latter may proceed through financial con-
straints, subsistence points, goods rationing or precautionary motives, but including each of these effects would
require substantial additional modeling efforts and result in an even larger model to estimate.
20
Evidence about the trading of government bonds from Germany in Zurich and Stockholm during the war
period is reported by Frey and Waldenström (2004), and shows that war events play a major role in volumes
traded. We take this as indirect evidence of our partial default channel.
17
The money growth spillover is not significantly different from zero. While inflation spikes during
war episodes in the data, our estimated model copes with this feature of the data by imposing
positive money growth innovations, especially at the end of war periods or in the immediate
post-war, rather than by producing a significantly positive money growth spillover.
Finally, comparing the baseline model with the restricted model (Table 8 in Appendix C) shows
that the baseline model produces a much larger marginal density on the US dataset. This
suggests that the baseline model is more strongly supported by the data. The Bayes factor is
e54.4 , a number showing strong evidence in favor of our baseline model. We conclude that the
shock on the war indicator variable and the associated transmission mechanisms help fit the
data. In addition, point estimates for the parameters that are estimated in both versions of the
model are in the same ballpark, showing that our baseline estimation is relatively robust.
5.2 Historical decompositions

We now analyze the contribution of the different shocks to the evolution of a subset of the
observed variables between 1901 and 1950, with a focus on war episodes. We simulate the model
only with the innovations of the shock considered and trace the evolution of the implied variables
of interest. Since the estimated model is linear, the observed time series are thus the sum of
the dynamics implied by each shock.21 We consider the contribution of the shock on the war
indicator variable ∆t , “supply shocks” (considering productivity and investment specific shocks
together), tax shocks (capital and labor income tax shocks together), the money growth, the
public spending shock and the foreign demand shock. Our variables of interest are the observed
growth rate of GDP, the observed inflation rate, the observed investment-to-GDP ratio and the
observed net-exports-to-GDP ratio. Figure 1 reports the historical decompositions that results
from the estimated smoothed structural innovations to the shocks (reported in Figure 17 in
Appendix E).
Starting with the observed dynamics of GDP growth, Figure 1 shows that the contribution of the
shock on the war indicator variable, that triggers the war specific regime for some parameters
and shocks spillovers, is relatively limited. It raises GDP growth on impact, mostly through
the spending spillover, but then produces a small contraction in GDP growth. It also contracts
GDP growth substantially the first year after the war period, again mostly because the positive
government spending spillover vanishes. Other shocks are crucial to actually match the dynamics
of GDP growth during war episodes. Most striking is the contribution of government spending
innovations, both during WWI and WWII, at the end of the conflicts, consistently with the
narratives of Section 2. These innovations come in addition to the spending spillover, that can
not account for GDP growth dynamics alone. Negative government spending innovations also
21
Up to the initial conditions and measurement error shock, that are disregarded to preserve the clarity of the
Figures.
18
Figure 1: US historical decompositions
(a) GDP growth (b) Inflation rate

0.4 0.25
0.2
War War
0.3
0.15
0.2 Money growth Money growth

0.1
0.05
0.1
TFP + investment TFP + investment
0
-0.05
Public spending Public spending
-0.1
-0.1
Foreign demand Foreign demand

-0.15
-0.2
-0.2
Taxes Taxes
-0.3 -0.25
0 5 10 15 20 25 30 35 40 45 50 0 5 10 15 20 25 30 35 40 45 50
19
(c) Investment to GDP (d) Net exports to GDP
0.1 0.08
War War
0.05 0.06
0 Money growth 0.04 Money growth
-0.05 0.02
-0.1 0
-0.15 -0.02
-0.2 -0.04
Taxes Taxes
-0.25 -0.06
0 5 10 15 20 25 30 35 40 45 50 0 5 10 15 20 25 30 35 40 45 50
contribute to the post-war slump by a substantial amount. Supply shock innovations contribute
only marginally to the dynamics of GDP growth during WWI but more substantially and pos-
itively during WWII. These effects are clearly consistent with the findings of McGrattan and
Ohanian (2010), who find that a large positive government spending and productivity shocks
account quite well for the dynamics of US GDP growth during WWII. Finally, our estimation
finds an important role for positive money growth innovation after WWII, to inflate some of the
war-related debt away. Overall, innovations to tax and foreign demand shocks play little role in
the dynamics of US GDP growth during war episodes.22
Other panels of Figure 1 suggest that the shock on the war indicator variable lowers inflation in
the US, lower the investment-to-GDP ratio and boosts net exports, mostly through the positive
trade spillover. Unless for net exports, however, this shock is not the most important historical
driver of these remaining variables during war episodes: inflation is more clearly driven by
innovations to the money growth rate and government spending shocks, the investment-to-GDP
ratio by innovations to the supply, government spending and tax shocks. Innovations to the
foreign demand shock also contribute quite substantially to the dynamics of net exports to GDP.
Overall, this historical decomposition sheds light on the contribution of the on the war indicator
variable to the US economy. It certainly drives some of the dynamics of macroeconomic aggre-
gates over these periods, but it is clearly not enough to fully account for those. In particular,
additional innovations to government spending shocks, productivity shocks and money growth
shocks matter at least as much.
5.3 Public spending multipliers

Our previous analysis makes clear that the rise in government spending during war episodes
is a robust feature of war periods in the US. On top of the spillover implied by the shock
on the war indicator variable, our historical decomposition suggests that a large fraction of
macroeconomic fluctuations during war episodes results from the rise of government spending. An
interesting question is thus: how do war episodes affect the value of public spending multipliers?
As explained in Appendix E, a shock to the war indicator variable already produces a very
large rise in public investment and consumption, that boosts output. Does an additional public
spending shock produce a larger or a lower spending multipliers in this case, compared to a shock
that would hit at the steady state? To answer this question, we draw values of our estimated
parameters from the posterior distributions and for each draw, compute the present-value public
consumption output multiplier when the shock hits in the first year of a war episode – imposing
that our war indicator variable is one during 5 years – and when the shock hits at the steady
state – when the war indicator variable is always zero. We conduct non-linear simulations under
22
As a by product of our analysis, our model finds that the Great Depression of the 30’s is mostly explained by
negative supply and money growth shocks, in line with most contributions on the subject (see Appendix H).
20
perfect foresight and disregard any kind of other source of uncertainty.23 The main purpose of
non-linear perfect foresight simulations is to account for the non-linearities implied by the shock
on the war indicator variable. The latter triggers a switch in parameters that crucially affect the
size of spending multipliers, such as the Edgeworth complementarity parameter or the capital
depreciation rate, with an effect on the dynamics of the real interest rate. The shock on the war
indicator variable also activates spillovers with persistent effects.24
Figure 2 below reports the value of the present-value multipliers, representing the discounted
dollar increase in output, consumption or investment that results from a dollar increase in public
consumption gt . It also reports multipliers confidence intervals. When the shock hits at the
steady state, the reference path of variable x against which deviations are computed is the
steady state. When the shock hits conditional on a 5-years war episode, the reference path of
variable x is the path implied by a shock on the war indicator variable only.25
Our results suggest that output present-value multipliers (PVMs hereafter) are rather large on
impact (larger than 1) and then fall under the combined effects of public spending and tax
rules that ensure debt sustainability in the long run. Long-run median present-value output
multipliers remain slightly above one. In addition, they are a little higher during war episodes
than during normal times. This difference essentially reflects the difference in the responses of
consumption and investment. During normal times, private consumption is crowded-in – due
to the Edgeworth complementarity between public and private goods in the utility function
of households – and investment increases slightly. During a war episode, private consumption
is crowded-out, and investment increases more substantially, producing overall larger median
estimates of the spending output multiplier. Yet, according to our estimates and confidence
intervals, this difference is not statistically significant. The key difference between war times and
normal times is that war episodes are characterized by a positive labor income tax spillover and a
negative capital income tax spillover: the latter alters the intertemporal price of consumption and
provides incentives for households to skew their current expenditure towards capital accumulation
against consumption. Hence, war episodes favor a larger positive response of investment and a
23
All the remaining innovations to shocks are assumed to be zero.
24
When estimated spillover parameters are not statistically different from zero, that is, when their 90% confidence
interval is wide and comprises positive and negative values, we simply impose the spillover parameter to be zero in
the simulations. This avoids “polluting” the simulations with spillover effects that are not informative according
to the estimation.
25
Let PT j
xgt+j − x

ss j=1 β
P V Mx,T = PT , for x = y, c, i (26)
j gg

j=1 β t+j − g
denote the multiplier in normal times and

PT j
xwar+g − xwar

war j=1 β t+j t+j
PV Mx,T = PT war+g war
, for x = y, c, i (27)
j − gt+j
j=1 β gt+j
denote the multiplier conditional on a 5-years war episode.
21
Figure 2: Present-value multipliers at different horizons
Black: normal times. Red: war episodes. Shaded areas correspond to one
standard deviation confidence intervals (16th and 84th percentiles) based on 120
replications of our non-linear perfect foresight simulations. Each replication
draws a vector of the estimated parameters from the posterior distributions and
computes the multipliers.
crowding-out of private consumption.
Our estimates of multipliers based on US data belong to the upper range of estimated spending
output multipliers reported by Ramey (2019). In addition, the question of state-dependent
multipliers has recently been addressed with two different set of conclusions. Auerbach and
Gorodnichenko (2012) find that multipliers are larger during periods of economic slack while
Owyang, Ramey and Zubairy (2013) find no particular difference for the US but a significant one
for Canada. In our model, multipliers differ during war episodes because when the war indicator
variable is one, Edgeworth parameters may differ and some war-specific effects are triggered
(spillovers, draft, default and capital depreciation), which can result in different transmission
mechanisms for public spending shocks. While this non-linearity seems to matter for the impact
response of consumption and investment, it does not produce significantly different output PVMs
at medium or long horizons. Our results thus seem to back those of Owyang et al. (2013) more
than those of Auerbach and Gorodnichenko (2012), based on our estimation on US data.
22
6 Results for UK, France and Germany
6.1 Estimation results
As in the case of the US, prior and posterior distributions are reported in Appendix D and
show that the estimation procedure provides enough information to update prior distributions.
Table 3 below is an extraction of the posterior means and 90% confidence intervals of some key
parameters for all countries, including the US for comparison purposes.
Table 3: Key parameter mean estimates, with 90% confidence intervals

France Germany UK US
Structural parameters
Calvo prices (η p ) 0.20 0.22 0.25 0.40
[0.15,0.25] [0.16,0.27] [0.20,0.31] [0.36,0.44]
w
Calvo wages (η ) 0.27 0.40 0.14 0.20
[0.14,0.39] [0.33,0.47] [0.05,0.22] [0.11,0.30]
Edgeworth comp. (νnorm ) 0.50 0.56 0.60 0.55
[0.47,0.53] [0.52,0.60] [0.59,0.62] [0.53,0.57]
Edgeworth comp. (νwar ) 0.35 0.48 0.46 0.51
[0.30,0.40] [0.40,0.55] [0.42,0.51] [0.47,0.55]
`
War draft (p ) 0.08 0.08 0.11 0.06
[0.02,0.14] [0.02,0.15] [0.03,0.18] [0.02,0.11]
b
War default (p ) 0.08 0.10 0.06 0.12
[0.03,0.13] [0.02,0.16] [0.02,0.10] [0.06,0.18]
δ
War capital depreciation (p ) 0.09 0.10 0.09 0.09
[0.02,0.15] [0.02,0.17] [0.02,0.15] [0.02,0.16]
Investment adj. cost (φi ) 4.04 4.68 3.74 4.27
[3.21,4.89] [3.85,5.52] [2.92,4.53] [3.49,4.05]
Trade elasticity (µ) 1.48 1.24 1.17 1.26
[1.31,1.65] [1.08,1.40] [1.08,1.26] [1.17,1.35]
Labor tax rule (dτ n ) 0.09 0.32 0.07 0.13
[0.07,0.12] [0.18,0.44] [0.06,0.09] [0.08,0.17]
Capital tax rule (dτ n ) 0.13 0.15 0.11 0.17
[0.08,0.18] [0.08,0.22] [0.08,0.15] [0.08,0.25]
Spending rule (dg ) 0.15 0.10 0.11 0.10
[0.09,0.21] [0.07,0.13] [0.07,0.15] [0.07,0.13]
War spillovers
Productivity (φa ) −0.05 −0.06 −0.02 0.01
[−0.08,−0.02] [−0.09,−0.03] [−0.04,−0.01] [−0.01,0.02]
Spending (φg ) 0.01 0.03 0.34 0.14
[−0.09,0.11] [−0.10,0.17] [0.22,0.47] [0.07,0.21]
Investment (φς ) −0.62 −0.46 −0.34 0.06
[−1.08,−0.17] [−0.91,−0.02] [−0.64,−0.05] [−0.31,0.41]
Trade (φy∗ ) −0.70 −0.08 −0.38 0.16
[−0.94,−0.45] [−0.29,−0.12] [−0.52,−0.24] [0.06,0.26]
Labor tax (φτ n ) −0.18 0.44 −0.14 0.13
[−0.25,−0.10] [0.24,0.63] [−0.20,−0.09] [0.06,0.20]
Capital tax (φτ n ) 0.38 −1.75 0.22 −0.34
[0.18,0.59] [−2.49,−1.00] [0.08,0.36] [−0.54,−0.14]
Money growth (φµ ) 0.03 −0.01 0.02 −0.02
[−0.03,0.10] [−0.07,0.05] [−0.00,0.05] [−0.04,0.01]
Notes: Results are extracted from Tables 5 to 8 in Appendix C.
Table 3 shows that countries other than the US have rather less sticky prices and more sticky
wages. According to our estimates, Germany features more overall stickiness than France or the
UK. In addition, France, Germany and the UK all feature significantly different estimates for
the Edgeworth parameter ν during normal times and during war times, public and private goods
being more strongly complementary during war episodes. This could result in different present-
value multipliers compared to the US. The point estimates for the war draft parameters are all
23
larger than in the US, the war partial default parameters are somehow smaller, and the war
capital depreciation parameters are roughly the same than in the US. The largest war-specific
depreciation factor is estimated at 10% for Germany, the country of our sample that arguably
suffered the largest war-related capital destruction, especially at the end of WWII.
Another interesting difference is the sign and size of war spillovers. As opposed to the US,
European countries of our sample feature negative productivity and investment war spillovers.
We find that these spillover parameters are not significantly different from zero for the US.
Similarly, a key difference is that the foreign demand spillover is found to be significant and
negative for European countries, and significant but positive for the US. Only the UK features a
positive public spending spillover while this parameter is muted for France and Germany. Finally,
France and the UK are characterized by a financing of war efforts through the capital income tax
along with a lower labor income tax, while Germany exhibits a financing pattern that is more
similar to the US: an increase in the labor income tax and a lower capital income tax. For all
countries, the spillover on money growth is muted.
Overall, our estimations find that many war-specific parameters are statistically different from
zero. This tends to comfort our theoretical assumptions. However, those war-related assump-
tions, triggered by the shock on the war indicator variable, do not capture completely the macroe-
conomic dynamics observed during war episodes, and the estimation has to resort to additional
innovations to structural shocks. For example, money spillovers are statistically non-significant
while we know that governments have used money growth to partly inflate the debt away. In
this respect, our model is able to capture this through two channels: the money growth spillover
or innovations to the money growth rule. If the behavior of the government is different during
WWI and WWII, or if positive money growth came only at the end of the conflicts and after,
our model may not capture a significant spillover and rather capture seigniorage through money
growth innovations. The same reasoning may apply for other spillovers, such as the public spend-
ing spillover in France and Germany, or the productivity spillover in the US. The former is found
to be non-significant while data suggest that public spending increased during both WWI and
WWII in France and Germany. The latter is muted while productivity shocks have been shown
to be important during WWII (see McGrattan and Ohanian (2010)). Our estimations find these
spillovers to be statistically non-significant but need positive innovations to the related structural
shocks (public spending or productivity) to fit the data.
As a final piece of evidence in favor of our assumptions concerning the way war episodes affect
the economy, Tables 5 to 7 in Appendix C all show that the marginal densities associated to
estimations of the baseline model are larger than those associated with the estimations of the
restricted model. Bayes factors are respectively e11.02 for France, e9.35 for Germany and e34.96 for
the UK, showing stronger support in favor of the baseline model. For European countries, as for
the US, the estimated values of parameters estimated in both models are also roughly similar.
24
6.2 Historical decompositions
We now proceed to the analysis of historical decompositions of key macroeconomic variables for
European countries. Figures 22 to 24 in Appendix G report those decompositions.
Starting with France, Figure 22 shows that the shock on the war indicator variable plays a much
bigger role in output dynamics than it does for the US. It explains the bulk of GDP growth
fluctuations, along with public positive spending and negative innovations to productivity and
investment shocks. A large part of the positive inflation dynamics during world wars is also
explained by this shock, especially during WWI. The early dynamics of inflation during WWII
are also well captured by the shock on the war indicator variable, but the late dynamics rely on
large positive money growth innovations to fit the data. The shock on ∆t lowers the investment-
to-GDP ratio and the net exports-to-GDP ratio and contributes to explain a large fraction of
those variables, especially during WWI. During the interwar period, the investment ratio is
almost exclusively driven by negative productivity and investment shocks, and the dynamics of
net exports is also driven by large foreign demand innovations. Figure 22 thus suggests that
the shock on the war indicator variable and the associated transmission mechanisms are more
important in explaining macroeconomic time series in France during world wars than it is for
the US.
Now looking at the historical decomposition for Germany, Figure 23 also shows that the shock on
the war indicator variable explains a substantial part of the observed GDP growth, the investment
ratio and, to a lesser extent, net exports during war episodes. The shock depresses GDP growth,
raises inflation, lowers the investment ratio and drives net exports in negative territories. During
world wars, the public spending shock and the productivity/investment shocks also contribute
to explain the dynamics of GDP growth and the investment ratio. The money growth shock
is crucial in accounting for the dynamics of inflation (especially during WWI), and the foreign
demand shock is key to the dynamics of net exports.26 In the case of Germany as well, the shock
on the war indicator variable plays a crucial role in explaining the dynamics of the economy
during world war episodes.
Let us finally look at the historical decomposition for the UK, in Figure 24. The results are more
contrasted, somewhere in between the US and France/Germany. On the one hand, the shock
on the war indicator variable contributes moderately and negatively to GDP growth dynamics,
and little to inflation during WWI but more during WWII. On the other hand, it contributes
quite significantly to the dynamics of the investment ratio and net exports, both during WWI
and WWII. Government spending shocks contribute massively to the dynamics of GDP growth,
of inflation during WWI and of the investment ratio during both world war episodes, and the
26
Notice that the hyperinflation of the early 20’s is captured by a sequence of large and positive money growth
shocks, that raise inflation, the investment ratio and net exports to GDP (see also Appendix H). The depression
of the 30’s appears in the form of negative investment and productivity shocks, as in the US.
25
foreign demand shock accounts for a substantial fraction of the dynamics of net exports during
WWII.
Our main conclusion from this exercise is that the shock on the war indicator variable accounts
for a relatively marginal fraction of observed variables for the US, helps fit some of the data for
the UK, and plays a critical role for France and Germany. We see this set of results as a useful
complement to the literature: while the effects of war episodes can be accounted for by large
spending and productivity shocks in the US and to a lesser extent in the UK, it is not the case
for other countries such as France or Germany. The “macroeconomics of war and peace” should
therefore allow for shocks such as our shock on the war indicator variable, that destroy part
of the stock of capital, affect the labor force, alter the dynamics of public debt, disrupt trade,
among other things, at least for countries such as France and Germany.
6.3 The macroeconomic effects of a simulated war episode

What are the effects of an artificial shock on the war indicator variable, i.e. a shock on ∆t ,
according to our estimated model? We consider the economy to be initially in the steady-
state and abstract from all other sources of uncertainty. We feed in a 5-year shock on ∆t and
solve the model using non-linear perfect foresight simulations.27 Figures 3 and 4 depict the
resulting macroeconomic effects in the three countries France, Germany and the UK, as well as
the dynamics for the US, for a comparison. While the historical decompositions made it clear
that the shock on the war indicator variable was not the sole (and for some countries like the
US even not the main) driver of the economy, this exercise reveals key differences implied by our
estimation concerning the conditional effects of a war episode.
According to Figures 3 and 4, a war episode in the UK has remarkably similar effects compared
to the US. GDP rises on impact, consumption and investment fall, driven by the crowding-out
effect of the massive rise in public spending. Because the latter is larger in the UK than in the
US, public debt rises much above the levels reached by the US, but the global pattern is the
same. One difference comes from the pattern of tax spillovers in the UK and in the US. In the
UK, tax spillovers imply a fall in the labor income tax and a rise in the capital income tax. The
pattern is the opposite for the US. These differences mainly alter the responses of hours worked
and real wages but do not feed back much to other key macroeconomic variables. An additional
difference is the large trade deficit observed in the UK, driven by the negative trade spillover.
The US show a reduced trade deficit because the trade spillover is positive, mitigating the trade
deficit generated by the rise in public spending.
Figures 3 and 4 show that a simulated war episode produces very different adjustments for France
and Germany, compared to the US and the UK. Both countries show similar dynamics, although
27
In other words, we consider an “MIT” shock on ∆t .
26
Figure 3: Impulse Response Functions to a 5-year MIT shock on ∆t - Macro variables
GDP Consumption Investment
10 0
0 0
% dev.
% dev.
% dev.
-10 -20 -20
-20
-40
-30 -40
-60
5 10 15 20 25 5 10 15 20 25 5 10 15 20 25
Capital stock Hours Real wage
0 40
0
% dev.
% dev.
% dev.
20 -10
-20
0 -20
-30
-40 -20
5 10 15 20 25 5 10 15 20 25 5 10 15 20 25
Net exports to GDP Terms of trade Inflation
0 40 10
% dev.
% dev.
-5 20
0
%
-10 0
-20 -10
-15 -40 -20
5 10 15 20 25 5 10 15 20 25 5 10 15 20 25
Years Years Years
Black: France, Blue: Germany, Red: US, Dashed: UK
Figure 4: Impulse Response Functions to a 5-year MIT shock on ∆t - Policy variables

Debt-to-GDP Real balances
200
0
% dev.
150
%
-20
100
-40
50
5 10 15 20 25 5 10 15 20 25
Public invt Public consumption
8 150
% dev.
% dev.
6 100
4 50
2 0
-50
0
5 10 15 20 25 5 10 15 20 25
Labor tax Capital tax
50
60
0
Level
Level
40 -50
20 -100
5 10 15 20 25 5 10 15 20 25
Years Years
Black: France, Blue: Germany, Red: US, Dashed: UK
27
the magnitude of adjustments is larger for Germany. In those countries, a war episode massively
lowers GDP – by almost 40% in France and more than 40% in Germany. Consumption falls more
than GDP, although public spending does not rise, suggesting that the depreciation of physical
capital and the draft play much more prominent roles than in the UK or the US. France displays
a massive trade deficit that reaches almost 20% of GDP, while Germany experiences a moderate
trade deficit. Because nominal rigidities are estimated to be larger in Germany, inflation rises
massively, by more than 20%, after reversing when the shock disappears. France and Germany
are characterized by opposite patterns regarding taxes spillovers: France sees a rise in the capital
tax and a fall in the labor tax (as the UK) and Germany sees a rise in the labor tax and a fall
in the capital tax (as the US).
This exercise reveals that a shock on the war indicator variable, such as estimated by our model,
has very different implications for the dynamics of the economy whether looking at the US or the
UK, and countries such as France or Germany. For the latter, unsurprisingly, the dynamics is
extremely negative for GDP growth, consumption, investment, inflation and net exports. We cer-
tainly do not claim that war episodes benefited to countries like the US or the UK, war obviously
had negative consequences for all belligerents. Our main point is that the consequences of war
episodes are quite different when those are not modeled as large spending and/or productivity
shocks, as it is usually done in a literature that focuses on the US and the UK.
6.4 Public spending output multipliers

We perform the same computation of public spending multipliers for France, Germany and the
UK. We proceed as in Section 4.3 and consider the effects of a shock on public consumption
during normal times and conditional on a 5-years war episode. Figure 5 reports the public
spending multipliers for the three countries, plus the US for a comparison, along with confidence
bands.
First, the levels of multipliers differ across countries. France has the largest multipliers, between
1.4 and 1.7 on impact, and the long-term multipliers are above one. Then, the US also fea-
ture large impact multipliers, between 1 and 1.3, and the long-term multipliers are above one.
Germany and the UK have lower impact multipliers, around 1, and much lower multipliers in
the long run (around 0.5 for the UK and 0 for Germany). For Germany, these lower long-run
multipliers are due to the very large estimated coefficient of the labor income tax feedback rule.
For the UK, the larger estimated Edgeworth parameter – signalling less complementarity – is
responsible, especially for normal-times multipliers.
Second, while Germany and the US have broadly similar impact and short-run output multipliers
during war episodes and during normal times, France and the UK feature larger multipliers during
war episodes. These differences are statistically significant, and suggest that, for those countries,
28
Figure 5: Present-value multipliers at different horizons
(a) France (b) Germany
(c) UK (d) US
Black: normal times. Red: War episodes. Shaded areas correspond to one standard deviation confidence intervals
(16th and 84th percentiles) based on 120 replications of our non-linear simulations. Each replication draws a vector
of the estimated parameters from the posterior distributions reported in Appendix D and computes PVMs.
29
the values of spending multipliers identified with war episodes might not be informative about
the value of multipliers during normal times. The most obvious driver of this result for the UK
and France is the larger difference in estimated Edgeworth parameters during normal times and
during war episodes: in both countries the Edgeworth parameter is significantly lower during
war episodes, making private consumption rise more after a spending shock than during normal
times, leading to larger output multipliers.
Overall, our conclusion is that war episodes are indeed periods during which government spending
rise massively. While interesting from the perspective of the identification of exogenous spending
shocks, the resulting output multipliers may not necessarily be interpreted as informative about
the size of multipliers during normal times because the structure of the economy may differ.
Our results suggest that war episodes are associated with large disruptions in the functioning
of the economy, with changes in the macroeconomic environment (households’ behavior, taxes,
capital efficiency) and hence with changes in transmission mechanisms, that might affect the size
of spending multipliers. The above exercise indicates that it would be more likely the case for
France and the UK, and less for Germany or the US.
7 Welfare losses from war episodes

Finally, we proceed to a welfare analysis. Using the series of smoothed structural shocks, i.e.
the series of shocks that best replicate our datasets, we simulate the model for each country
using estimated parameters.28 First, we feed the model with all smoothed structural shocks.
Then, we run a counterfactual exercise and feed the model with all the smoothed structural
shocks without the shock on the war indicator variable. Then we compute the welfare losses
induced by war episodes, measured as the Hicksian consumption equivalent, i.e. the consumption
compensation that makes households indifferent between living in a world with war episodes and
in a world without them. This approach is also used to measure the welfare costs of other sources
of macroeconomic fluctuations. Finally, we report the total welfare losses from fluctuations,
measured as the Hicksian equivalent that makes households indifferent between an economy with
fluctuations and an economy that remains in the “peaceful” steady state.29 The results are
reported in Table 4.
Table 4 suggests that war episodes have large negative welfare effects for France and Germany,
moderately negative effects for the UK and positive effects in the US. These results can be
understood through the behavior of key macroeconomic variables during war episodes. France
and Germany experience large drops in private consumption that are not compensated by a
rise in public spending – remember, the spending spillover is muted for those countries. They
also experience large hikes in hours worked. Both lower consumption paths and higher paths
28
We only abstract from innovations in the measurement error shock.
29
As usual when introducing money in the utility function, the welfare effects of real balances are simply ignored.
30
Table 4: Welfare costs, in percents
Welfare cost of: France Germany UK US
War episodes 0.6892 0.8752 0.1185 −0.0382
Productivity and investment shocks −0.6391 −0.6607 −0.0196 −0.0350
Foreign demand shocks 0.0602 0.0074 0.0061 0.0013
Public spending shocks −0.1991 −0.0334 0.2768 −1.5256
Tax shocks −0.5378 0.0049 −0.0363 −0.2480
Money growth shocks 0.0324 −0.0245 0.0054 0.0022
All fluctuations 0.0153 0.0103 0.0074 0.0586
for hours worked generate welfare losses. Quantitatively speaking, Germany suffers the largest
welfare losses, representing 0.88% of permanent consumption. For France, the welfare losses are
of the same order of magnitude, around 0.69% of permanent consumption.
For the UK and the US, the negative effects of war episodes are partly compensated by large
jumps in public spending that supports the economy and brings welfare gains to the households
since (i ) public consumption is a direct argument of the utility function and (ii ) private con-
sumption is crowded-in when public consumption increases, which attenuates the fall in private
consumption. For the UK, these effects attenuate the welfare losses from war episodes, that
reach 0.12% of permanent consumption, lower than those found for France and Germany. For
the US, the “positive” effects on utility are large enough to generate small welfare gains, around
0.04% of permanent consumption.30
As robustness check, we also compute the total welfare losses from fluctuations. Based on the
whole sample, the latter are much smaller, because the negative welfare effects of some shocks are
almost fully compensated by the positive effects of other shocks. Overall, the resulting welfare
losses from fluctuations are very small, in the order of magnitude typically reported by Lucas
(1987) and the subsequent literature, ranging from 0.01% to 0.06% of permanent consumption.
8 Conclusion
War episodes induce large shifts in macroeconomic variables provoked by a series of potential
related causes: destruction, military draft, changes in national borders, large wealth transfers
from a defeated country to a victorious one, interventions of governments that raise public
spending, inflate the public and military sector, alter the structure of the labor market and
impose restrictions on financial and good markets.
This paper proposed an estimated open economy model that would be suited to capture the
dynamics of countries such as France, Germany, the UK and the US both during normal times
30
This result should be interpreted with caution. Clearly, US households would have preferred not to be involved
in world war episodes, as the costs of those world war episodes are immensely larger than just their macroeconomic
costs.
31
and during war episodes. Of course, all the above aspects could not be fully taken into account
but some of them (capital destruction, draft, partial default on public debt) were, and were
shown to be statistically significant. In addition, we shown that the macroeconomic effects of
war episodes were not fully accounted for by the shock on the war indicator variable, and needed
to be completed with other shocks. A massive increase in public spending (UK, US), negative
productivity and investment shocks (France, Germany and the UK), a negative foreign demand
shock (France and the UK), and variations in the labor and capital income tax rates, were also
important to fit the data. Our main conclusions thus complement Braun and McGrattan (1993)
and McGrattan and Ohanian (2010) about the key role of public spending shocks on the US and
UK economy during world wars episodes. For Continental Europe, however, war episodes affected
the economy through the channels of capital destruction, military draft, partial default on public
debt, and spillovers. In these countries the public spending spillover was muted, investment,
productivity and trade spillovers were negative and significant.
Finally, our estimation results were used investigate the size and state-dependence of public
spending output multipliers and about the welfare losses from war episodes. Some countries
(Germany and the US) had statistically similar patterns of spending output multipliers during
normal times and during war episodes, while some others (France and the UK) featured larger
multipliers during war episodes. The welfare analysis showed that Germany and France were
the countries of our sample that suffered most in terms of the welfare effects of macroeconomic
dynamics, while the welfare (economic) costs were smaller for the US and the UK.
References
Ambler, S. and A. Paquet, Stochastic Depreciation and the Business Cycle, International Eco-
nomic Review, 1994, 35 (1), 101–116.
An, S. and F. Schorfheide, Bayesian Analysis of DSGE Models, Econometric Reviews, 2007, 26
(2-4), 113–172.
Anderton, C. H. and J. R. Carter, The Impact of War on Trade: An Interrupted Times-Series

Study, Journal of Peace Research, 2001, 38 (4), 445–457.
Auerbach, A. J. and Y. Gorodnichenko, Fiscal Multipliers in Recession and Expansion, in Fiscal

Policy after the Financial Crisis NBER Chapters, National Bureau of Economic Research, Inc,
2012, pp. 63–98.
Auray, S., A. Eyquem, and F. Jouneau-Sion, Wars and Capital Destruction, Journal of Economic
Dynamics and Control, 2014, 41 (C), 224–240.
, , and P. Gomme, A Tale of Tax Policies in Open Economies, International Economic

Review, 2016, 57 (4), 1299–1333.
32
Backus, D. K., P. J. Kehoe, and F. E. Kydland, International Real Business Cycles: Theory vs.
Evidence, Federal Reserve Bank of Minneapolis Quarterly Review, 1993, 17, 14–29.
Barro, R. J. and J. F. Ursúa, Stock-market Crashes and Depressions, Research in Economics,

2017, 71 (3), 384–398. Special issue on Macroeconomics.
Ben Zeev, N. and E. Pappa, Chronicle of a War Foretold: The Macroeconomic Effects of Antic-
ipated Defense Spending Shocks, Economic Journal, 2015, 127 (603), 1568–1597.
Bouakez, H. and N. Rebei, Why Does Private Consumption Rise after a Government Spending
Shock?, Canadian Journal of Economics, 2007, 40 (3), 954–979.
Braun, A. R. and E. R. McGrattan, The Macroeconomics of War and Peace, in NBER Macroe-
conomics Annual 1993, Volume 8 NBER Chapters, National Bureau of Economic Research,
Inc, 1993, pp. 197–258.
Canova, F., F. Ferroni, and C. Matthes, Detecting and Analyzing the Effects of Time-Varying
Parameters in DSGE Models, International Economic Review, 2019. Forthcoming.
Chari, V. V., Patrick J. Kehoe, and Ellen R. McGrattan, Business Cycle Accounting, Economet-
rica, 2007, 75 (3), 781–836.
Devereux, M. B. and G. W. Smith, Transfer Problem Dynamics: Macroeconomics of the Franco-

Prussian War Indemnity, Journal of Monetary Economics, 2007, 54 (8), 2375–2398.
Fishback, P., US monetary and fiscal policy in the 1930s, Oxford Review of Economic Policy,
2010, 26 (3), 385–413.
Flandreau, M. and F. Zumer, The Making of Global Finance 1880-1913, OECD Development
Centre Studies, 2004. 144 pages.
Frey, B. S. and D. Waldenström, Markets Work in War: World War II Reflected in the Zurich
and Stockholm Bond Markets, Financial History Review, 2004, 11 (01), 51–67.
Friedman, M. and A. J. Schwartz, A Monetary History of the United States, 1867-1960 NBER
Books, National Bureau of Economic Research, Inc, July 1963.
Furlanetto, F. and M. Seneca, New Perspectives on Capital Depreciation Shocks as Sources of

Business Cycle Fluctuations, Macroeconomic Dynamics, 2014, 18 (6), 1209–1233.
Ghironi, F. and M. J. Melitz, International Trade and Macroeconomic Dynamics with Heteroge-
neous Firms, The Quarterly Journal of Economics, 2005, 120 (3), 865–915.
Glick, R. and A. M. Taylor, Collateral Damage: Trade Disruption and the Economic Impact of
War, The Review of Economics and Statistics, 2010, 92 (1), 102–127.
33
Justiniano, A., G. Primiceri, and A. Tambalotti, Investment Shocks and the Relative Price of
Investment, Review of Economic Dynamics, 2011, 14 (1), 101–121.
Leeper, E. M., Equilibria under ’Active’ and ’Passive’ Monetary and Fiscal Policies, Journal of
Monetary Economics, 1991, 27 (1), 129–147.
, N. Traum, and T. B. Walker, Clearing Up the Fiscal Multiplier Morass, American Economic
Review, 2017, 107 (8), 2409–2454.
Lucas, R. E., Models of business cycles, Oxford: Blackwell, 1987.
Martin, P., T. Mayer, and M. Thoenig, Make Trade not War?, Review of Economic Studies,
2008, 75 (3), 865–900.
McGrattan, E. R. and L. E. Ohanian, Does Neoclassical Theory Account for the Effect of Big
Fiscal Shocks? Evidence from WWII, International Economic Review, 2010, 51 (2), 509–532.
Mitchell, B. R., International historical statistics: the Americas, 1750-1993. 1998. London:
Macmillan Reference / New York: Stockton Press.
Monnet, C. and W. E. Weber, Money and interest rates, Quarterly Review, 2001, (Fall), 2–13.
Ohanian, L. E., The Macroeconomic Effects of War Finance in the United States: World War II
and the Korean War, American Economic Review, 1997, 87 (1), 23–40.
Owyang, M. T., V. A. Ramey, and S. Zubairy, Are Government Spending Multipliers Greater
during Periods of Slack? Evidence from Twentieth-Century Historical Data, American Eco-
nomic Review, 2013, 103 (3), 129–34.
Piketty, T. and G. Zucman, Capital is Back: Wealth-Income Ratios in Rich Countries, 1700-2010,
Quarterly Journal of Economics, 2014, 129 (3), 1255–1310.
Ramey, V. A., Can Government Purchases Stimulate the Economy?, Journal of Economic Lit-
erature, 2011, 49 (3), 673–85.
, Identifying Government Spending Shocks : It’s All in the Timing, Quarterly Journal of
Economics, 2011, 126 (1), 1–50.
, Ten Years After the Financial Crisis: What Have We Learned from the Renaissance in Fiscal
Research?, The Journal of Economic Perspectives, 2019, 33 (2), 89–114.
and M. Shapiro, Costly Capital Reallocation and the Effects of Government Spending,
Carnegie-Rochester Conference Series on Public Policy, 1997, 48 (1), 145–194.
Siu, H. E., The Fiscal Role of Conscription in the U.S. World War II Effort, Journal of Monetary
Economics, 2008, 55 (6), 1094–1112.
34
Smets, F. and R. Wouters, An Estimated Dynamic Stochastic General Equilibrium Model of the
Euro Area, Journal of the European Economic Association, 2003, 1 (5), 1123–1175.
and , Comparing Shocks and Frictions in US and Euro Area Business Cycles: A Bayesian
DSGE Approach, Journal of Applied Econometrics, 2005, 20 (2), 161–183.
Young, A. and J. C. Musgrave, Estimation of Capital Stock in the United States, in The Mea-
surement of Capital NBER Chapters, National Bureau of Economic Research, Inc, March
1980, pp. 23–82.
35
A Datasets
Figure 6: France
0.2 0.2
0.05
0.1
0.1
0 0
0
-0.1
-0.1 -0.05
-0.2
-0.2 -0.3
-0.1
1900 1950 2000 1900 1950 2000 1900 1950 2000
0.3
0 0.3
0.2 0.2
-0.1 0.1
0.1
0
0 -0.2 -0.1
-0.2
1900 1950 2000 1900 1950 2000 1900 1950 2000
0.4 0.4 1
0.8
0.2 0.2
0.6
0 0
0.4
-0.2 -0.2 0.2
0
1900 1950 2000 1900 1950 2000 1900 1950 2000
36
Figure 7: Germany
0 0.05
0.1
-0.2 0
-0.4 0 -0.05
-0.6
-0.1 -0.1
-0.8
-1 -0.15
-0.2
1900 1950 2000 1900 1950 2000 1900 1950 2000
0.05 1
0.15
0.1 0.8
0.05 0.6
0
0 0.4
-0.05 0.2
-0.1 -0.05 0
1900 1950 2000 1900 1950 2000 1900 1950 2000
0.8 1
0.6 0.8
0
0.6
0.4
-1
0.4
0.2
-2 0.2
0
0
1900 1950 2000 1900 1950 2000 1900 1950 2000
37
Figure 8: United Kingdom
0.1
0.05
0.05 0
0
0
-0.05
-0.05
-0.05
-0.1
-0.1 -0.1
1900 1950 2000 1900 1950 2000 1900 1950 2000
0.05
0.15
0.1
0.1 0
0.05 0
0 -0.05
-0.1
-0.05
-0.1
1900 1950 2000 1900 1950 2000 1900 1950 2000
1
0.4 0.3
0.8
0.2
0.1 0.6
0.2
0 0.4
0 -0.1
0.2
-0.2
0
1900 1950 2000 1900 1950 2000 1900 1950 2000
38
Figure 9: United States
0.05
0.1
0.05
0.05 0
0 0 -0.05
-0.05
-0.05 -0.1
-0.1
-0.1 -0.15
-0.15
1900 1950 2000 1900 1950 2000 1900 1950 2000
0.15 0.04 0.1

0.1
0.02
0.05 0
0
0
-0.1
-0.05 -0.02
1900 1950 2000 1900 1950 2000 1900 1950 2000
1
0.4 0.6
0.8
0.4
0.2 0.6
0.2
0.4
0 0
-0.2 0.2
-0.2 0
1900 1950 2000 1900 1950 2000 1900 1950 2000
39
B Model
B.1 Households
We consider a unit continuum of households indexed in j that maximize their lifetime welfare
(∞ !)
X
s−t cs (j)1−σc (ms (j) /ps )1−σm ns (j)1+ψ
Wt = Et + χm − χn (B.1)
e
β
s=t
1 − σc 1 − σm 1+ψ
In the welfare index, total consumption e

cs (j) is a bundle of individual consumption of the private
good ct (j) and the total consumption of public good gt
νt −1
νt −1 νt −1 ν
t
νt
ct (j) = ϕct (j) t + (1 − ϕ) gt
e ν , νt > 0 (B.2)
When νt > 1, private and public goods are substitutes while νt < 1 implies complementarity. We
introduce the possibility that this parameter takes different values whether countries experience
a war episode or not, that is
νt = (1 − ∆t ) νnorm + ∆t νwar (B.3)
where ∆t = {0, 1} is an indicator variable that equals one during war episodes and zero in
normal times. Regarding other variables of the welfare index, mt (j) /pt is the level of real
money balances, and nt (j) is the total amount of labor supply, expressed in hours. Parameters
σc and σm are the degrees of risk-aversion with respect to total consumption and real money
balances, and ψ is the inverse of the Frisch elasticity of labor supply.
The budget constraint of agent j is

et ft (j) + bt (j) + mt (j) + pt ct (j) + it (j) + acft (j) = et r∗ ft−1 (j) + rt−1 1 − ∆t pb bt−1 (j)

+ mt−1 (j) + 1 − τtk rtk zt (j) − pt aczt (j) + pt τtk δt kt−1 (j)
+ (1 − τtn ) wt (j) nt (j) + Πt (j) + taxt (j) (B.4)
In this equation, et is the nominal exchange rate, ft (j) denotes the nominal value of foreign
bonds returning a constant risk-free rate r∗ between period t and t + 1 and bt (j) denotes the
nominal value of government bonds returning a nominal rate rt between period t and t + 1.
Government bonds can be subject to partial default that lowers the face value of the stock of
public debt by pb percents in the event of a war episode. Further, pt is the consumption price
level, it (j) is the investment in physical capital, τtn is the labor income tax rate, wt (j) is the
nominal wage rate paid to type-j labor, τtk the tax rate on capital income, that comes with a tax
deduction on depreciated capital, rtk is the gross nominal return on the capital stock, zt (j) is the
utilization rate of private capital, δt the time-varying depreciation rate of capital. Variable taxt
40
is a lump-sum transfer that is rebated to the households. Capital depreciation is an additional
channel through which war episodes can affect the dynamics of the economy as we assume

δt = δ 1 + ∆t pδ (B.5)
Finally, acft (j) = (φf /2) (et ft (j) /pt − ef (j) /p)2 is an adjustment cost on real net foreign assets,
aczt (j) = (φz /2) (zt (j) − 1)2 is a utilization rate adjustment cost and Πt (j) denotes the profits
from monopolistic firms paid to household j. An additional constraint to this optimization
problem is the law of capital accumulation
2 !
φi it (j)
kt (j) = (1 − δt ) kt−1 (j) + ζt 1− −1 it (j) (B.6)
2 it−1 (j)
The latter is subject to investment adjustment costs and to investment shocks ζt . Formally the
law of motion of investment shocks is:
ζt = (1 − ρζ ) + ρζ ζt−1 + φζ ∆t + εζt (B.7)
Households maximize the welfare function subject to the budget constraints with respect to
consumption, government and foreign bonds, as well as the quantity of money. First-order
conditions imply
( )
uc,t+1 (j) rt 1 − ∆t+1 pb
βEt = 1 (B.8)
uc,t (j) πt+1
et+1 r∗

uc,t+1 (j)
βEt = 1 (B.9)
uc,t (j) et πt+1 (1 + φf (ftr (j) − f r ))
( )
rt 1 − ∆t+1 pb − 1

um,t (j) − uc,t (j) Et = 0 (B.10)
rt (1 − ∆t+1 pb )
where πt = pt /pt−1 is the CPI inflation rate, uc,t (j) is the marginal utility of consumption,
um,t (j) the marginal utility of real money balances, and ftr (j) = et ft (j) /pt the real value of net
foreign assets. The different types of labor offered by households are imperfectly substitutable,
making them monopolistic wage-setters. Labor demands are
−θw
wt (j)
nt (j, ω) = nt (ω) (B.11)
wt
where θw is the elasticity of substitution between types of labor. Those demands are taken into
account by households along with the constraints when optimizing. Further, we assume sticky
wages whereby wage-setters face an individual probability 1 − η w to be allowed to re-optimize
and a probability η w to keep their previous-period nominal wage unchanged. The corresponding
41
optimal nominal wage wt (j) is thus
∞ w
X θ un,t+i (j) n wt (j)
(βη w )i Et nt+i (j) + (1 − τt ) =0 (B.12)
θw − 1 uc,t+i (j) pt+i
i=0
where un,t (j) is the marginal disutility of hours worked. The dynamics of nominal wages are
w w w
wt1−θ = η w (wt−1 )1−θ + (1 − η w ) wt (j)1−θ (B.13)
From now on we assume perfect risk-sharing among households and drop the j indices. Defining
qt pt λt as the Lagrange multiplier associated with the capital accumulation constraint, first-order
conditions with respect to the capital stock, investment, and utilization are respectively:
( !)
uc,t+1 rk
k t+1 k
βEt qt+1 (1 − δt+1 ) + 1 − τt+1 + τt+1 δt+1 = qt (B.14)
uc,t pt+1

φi 2 uc,t+1 2
qt ζt 1 − γt − φi γt (1 + γt ) + βEt qt+1 ζt+1 φi γt+1 (1 + γt+1 ) = 1 (B.15)
2 uc,t

1 − τtk rtk − φz (zt − 1) = 0 (B.16)
where γt = (it /it−1 ) − 1 is the growth rate of private investment. Consumption and capital goods
are made of domestic (d) and foreign (f ) goods. Adjustment costs are also paid in units of that
composite good. The latter is defined as:
µ−1 µ−1
µ
1 1 µ−1 n o
µ µ
xt = (1 − α) µ xd,t + α µ xf,t , x = c, i, acf , acz , aci (B.17)
where 1 − α is the degree of home bias and µ the elasticity of substitution between domestic and
foreign goods. The corresponding price index is
1
1−µ 1−µ 1−µ
pt = (1 − α) pd,t + α (et pf,t ) (B.18)
where pd,t and pf,t respectively denote the prices of domestic and foreign goods, expressed in
units of local currency, and where et is the nominal exchange rate. Optimal expenditure on each
good implies
pd,t −µ µ

1−µ 1−µ
xd,t = (1 − α) xt = (1 − α) 1 − α + αst xt (B.19)
pt
et pf,t −µ µ

µ−1 1−µ
xf,t = α xt = α (1 − α) st + α xt (B.20)
pt
for x = c, i, acf , acz , aci where st = et pf,t /pd,t stands for terms of trade.

42
B.2 Firms
There is a continuum of final good producers indexed in ω, with the following production function
1−ι
yt (ω) = at kts (ω)ι 1 − ∆t p` `t (ω) (B.21)
where kts is a measure of capital services used in production and 1 − ∆t p` `t is the amount of

labor that firms use to produce. Total labor `t is affected by a military draft that lowers the
amount of labor used by private firms during war episodes. Finally, at is a productivity measure
following an autoregressive process. Again, we introduce the possibility that war episodes have
indirect effects on productivity through channels that are not explicitly specified in the model:
at = (1 − ρa ) + ρa at−1 + φa ∆t + εat (B.22)
Cost minimization implies

ι wt
kts (ω) = `t (ω) (B.23)
1 − ι rtk
which can be used to derive an expression of the nominal marginal cost
ι
rtk (wt )1−ι (1 − ι) 1 − ∆t p` + ι

mct (ω) = mct = 1−ι
(B.24)
at ιι (1 − ι)(1−ι) (1 − ∆t p` )
The adjustment of production prices is also subject to Calvo contracts. Re-setters face the
following problem
∞
X
p i uc,t+i pt
max (βη ) Et (pd,t (ω) yt+i (ω) − mct+i yt+i (ω)) (B.25)
pd,t (ω) uc,t pt+i
i=0
taking into account the demand addressed to firm ω

−θp
pd,t (ω)
yd,t (ω) = yd,t (B.26)
pd,t
The optimal pricing condition is thus

∞
θp

X
p i uc,t+i
(βη ) Et yt+i (ω) pd,t (ω) − p mct+i =0 (B.27)
pt+i θ −1
i=0
while the dynamics of prices are given by
p p p
p1−θ
d,t = η p (pd,t−1 )1−θ + (1 − η p ) pd,t (ω)1−θ (B.28)
43
B.3 Monetary policy, government and aggregation
The government is in charge of both monetary policy, public spending and fiscal policy. Its budget
constraint is

bt + mt + τtn wt nt + τtk rtk zt − pt δt kt−1 + taxt

= rt−1 1 − ∆t pb bt−1 + mt−1 + ∆t p` wt `t + pd,t (gt + igt ) (B.29)
The left-hand side shows government revenues and the right-hand side features government
expenditure. In particular, the government has to pay drafted people. We assume that fiscal
solvency is achieved in the medium run by tax rules and a government spending rule. These
rules imply that policy instruments react to the deviations of real public debt to its steady state
value. In addition, we consider war-specific spillovers on taxes and spending. Finally, policy
instruments are also affected by autoregressive exogenous shocks:
gt = g + dτn (brt−1 − br) + φg ∆t + sgt (B.30)

τn
τtn = τ n + dτ n (brt−1 − br) + φτ n ∆t + st (B.31)
τk
τtk = τ k + dτ k (brt−1 − br) + φτ k ∆t + st (B.32)
where
sgt = ρg sgt−1 + εgt (B.33)

n
+ ετt n (B.34)
k
+ ετt k (B.35)
Finally, we consider that money supply grows at a rate µm

t = mt /mt−1 that follows an AR(1)
process augmented with a potential war-specific component:
µm m m
t = (1 − ρm ) + ρm µt−1 + φµ ∆t + εt (B.36)
Further, as in Auray et al. (2014), the government adjusts the amount of public investment to
secure a given constant level of public unproductive capital k g :
igt = δt k g (B.37)
An equilibrium of the model is as a path of endogenous variables that, conditional on the dy-
namics of exogenous variables, (i ) satisfies households and firms first-order conditions, and (ii )
clears markets. Equilibrium in the labor market implies
Z 1 Z 1Z 1
`t = `t (ω) dω = nt (j, ω) djdω = Υw
t nt (B.38)
0 0 0
44
R1 w
where Υw
t = 0 (wt (j) /wt )−θ dj is the dispersion of wages. Equilibrium in the market for
capital services gives Z 1
kts = kts (ω) dω = zt kt−1 (B.39)
0
and equilibrium in goods market implies

−µ −µ
pd,t pd,t
yt = (1 − α) ytd +α yt∗ + gt + it + igt (B.40)
pt p∗t

where ytd = ct + it + φf (ftr − f r )2 /2 + φz (zt − 1)2 /2 kt−1 and yt∗ is the foreign counterpart of
ytd . It is subject to autoregressive shocks, and augmented with a war-specific component:
∗
∗
yt∗ = (1 − ρy∗ ) y ∗ + ρy∗ yt−1 + φy∗ ∆t + εyt (B.41)
Finally, the aggregate production function is

1−ι
Υpt yt = at (zt kt−1 )ι Υw
t 1 − ∆t p`
nt (B.42)
R1 p
where Υpt = 0 (pd,t (ω) /pd,t )−θ dω is the dispersion of producer prices. This open economy
model is closed by the net foreign asset equation, obtained by consolidating households, firms
and government budget constraints
r∗ ft−1
r

et
ftr = + nxt (B.43)
et−1 πt
where nxt stands for net exports, defined as
pd,t φf r φz
nxt = (yt − gt − igt ) − ct − it − (f − f r )2 − (zt − 1)2 kt−1 (B.44)
pt 2 t 2
B.4 The steady state

We analyze the linearized dynamics of the model around the closed economy steady-state without
inflation, where we also assume s = 1. The steady-state of the economy is one where the war
indicator variable is zero, i.e. ∆ = 0. Further, by definition, γ = 0, and the dynamics of q gives
q = 1. The condition on the utilization rate thus gives
φz (z − 1)
rk = (B.45)
1 − τk
which plugged in the investment equation pins down z

s
2 (β −1 + δ (1 − τ k ) − 1)
z= 1+ (B.46)
φz
45
as well as rk . Next use the real marginal cost to get the real wage w
1
(θ − 1) ι −ι 1−ι
w= ι (1 − ι)(1−ι) rk (B.47)
θ
Further, we impose a = n = 1 and use
ι w
k= (B.48)
1 − ι rk z
and the aggregate production function

y = (zk)ι (B.49)
to get consumption
φz
c = 1 − κ − κi y − δk − (z − 1)2 k

(B.50)
2
where κ = g/y and κi = ig /y are the (imposed) steady-state share of public consumption and
investment expenditure in GDP. Finally, the steady-state wage-setting equation gives the value
of χn that normalizes n to one.
46
B.5 Set of estimated equations
Households
1−σ
−1 1−σc (mrt ) m n1+ψ
ut = (1 − σc ) ct + χm − χn t
1 − σm 1+ψ
e
νt −1 νt −1
νtν−1
t
ct
e = ϕct νt + (1 − ϕ) gt νt
νt −1
νt −1 νt −1
νtν−1 !1−σc νt −1 νt −1

t
νt νt νt
uc,t = ϕct ϕct + (1 − ϕ) gt / ct ϕct νt + (1 − ϕ) gt νt
−σm
um,t = χm (mrt )
un,t= −χn nψ
(t )
uc,t+1 rt 1 − ∆t+1 pb
1 = βEt
uc,t πt+1
det+1 r∗

uc,t+1
1 = βEt
uc,t πt+1 (1 + φf ftr )
( )
rt 1 − ∆t+1 pb − 1
um,t = uc,t Et
rt (1 − ∆t+1 pb )

uc,t+1 k
rk k

qt = βEt qt+1 (1 − δt+1 ) + 1 − τt+1 rt+1 + τt+1 δt+1
uc,t

φi 2 uc,t+1 2
1 = qt ζt 1 − γt − φi γt (1 + γt ) + βEt qt+1 ζt+1 φi γt+1 (1 + γt+1 )
2 uc,t
γt = it /it−1 − 1

φi
kt = (1 − δt ) kt−1 + ζt 1 − γt2 it
2
where det = et /et−1 and rtrk = rtk /pt .
Firms
rtrk φz (zt − 1) / 1 − τtk

=
ιwtr nt / (1 − ι) zt rtrk

kt =
ι 1−ι
rtrk (wtr ) (1 − ι) 1 − ∆t p` + ι
mcrt = (1−ι) 1−ι
at ιι (1 − ι) (1 − ∆t p` )
ι 1−ι
Υpt yt = at (zt kt−1 ) Υw
t 1 − ∆t p` nt
where wtr = wt /pt and mcrt = mct /pt
47
Wage and produce price inflation
θ w −1
1−θw w
(πtw ) = 1 − (1 − η w ) wt1 /wt2 /η
n w o
1+θ
wt1 = βη w Et wt+1 1 w
/πt+1 + (θw /(θw − 1))(−un,t /uc,t )/((1 − τtn ) wtr )

πt+1
n θ w o
wt2 = βη w Et wt+1 2 w
πt+1 /πt+1 + nt
w θ
w −θw
Υw
t = η w Υw t−1 (πt ) + (1 − η w ) wt1 /wt2
θp −1 1−θp p
πtd = 1 − (1 − η p ) p1t /p2t /η
n 1+θp o 1/(µ−1)
p1t = βη p Et p1t+1 πt+1 d
/πt+1 + (θp /(θp − 1))uc,t yt mcrt 1 − α + αs1−µ
t
n p o
θ
p2t = βη p Et p2t+1 πt+1 d

/πt+1 + uc,t yt
θ p −θp
Υpt = η p Υpt−1 (πtp ) + (1 − η p ) p1t /p2t
Consumer price inflation and terms of trade
1/(1−µ)
πt = πtd 1 − α + αs1−µ
t / 1 − α + αs1−µ
t−1
st /st−1 = det+1 π ∗ /πtd
where π ∗ = 1.
Goods market clearing, net foreign assets, net exports
µ/(1−µ)
2 2
yt = (1 − α) 1 − α + αs1−µ
t ct + it + φf (ftr ) /2 + φz (zt − 1) /2 kt−1
µ/(1−µ)
+ α α + (1 − α) s1−µ
t yt∗ + gt + it + igt
ftr = β −1 det ft−1

r
/πt + nxyt × yt
1/(µ−1)
1−µ g φf r 2 φz 2
nxyt = 1 − α + αst (yt − gt − it ) − ct − it − (f ) − (zt − 1) kt−1 /yt
2 t 2
Government
igt = δt k g = δt κi y/δ

mrt /mrt−1 = µm
t /πt
1/(1−µ)
brt + mrt + τtn − ∆t p` wtr nt + τtk rtrk zt − δt kt−1 1 − α + αs1−µ

t
= rt−1 1 − ∆t pb brt−1 /πtd − φb (brt − br ) + πt mrt−1 /πtd + gt + igt

gt = g + dτn (brt−1 − br) + φg ∆t + sgt

n
τtn = τ n + dτ n (brt−1 − br) + φτ n ∆t + sτt
k
τtk = τ k + dτ k (brt−1 − br) + φτ k ∆t + sτt
48
Switching parameters and shocks
νt = (1 − ∆t ) νnorm + ∆t νwar
= δ 1 + ∆ t pδ

δt
at = (1 − ρa ) + ρa at−1 + φa ∆t + εat
sgt = ρg sgt−1 + εgt
ζt = (1 − ρζ ) + ρζ ζt−1 + φζ ∆t + εζt
∗
yt∗ = (1 − ρy∗ ) y ∗ + ρy∗ yt−1
∗
+ φy∗ ∆t + εyt
n
+ ετt n
k
+ ετt k
µm
t = (1 − ρm ) + ρm µm m
t−1 + φµ ∆t + εt
yterr err
= ρerr yt−1 + εerr
t
Additional variables
1/(1−µ)
τtn − ∆t p` wtr nt + τtk rtrk zt − δt kt−1 1 − α + αs1−µ

Tt = t
cgyt = (ct + gt ) /yt

dyt = brt /yt
myt = mrt /yt
iyt = (it + igt ) /yt
Measurement equations
dytobs = log yt − log yt−1 + yterr

dcobs
t = log ct − log ct−1
πtobs = log πt−1 / πtd,obs = log πt−1
d
dbobs
t = log brt − log brt−1
dtobs
t = log Tt − log Tt−1
cgytobs = cgyt − cgy
nxytobs = nxyt − nxy
iytobs = iyt − iy
∆obs
t = ∆t
49
C Estimation results
Table 5: Country-specific estimated parameters for France

Priors Post. Baseline Post. Restricted
Dist. Mean Sd. Mean Inf. Sup. Mean Inf. Sup.
Calvo prices (η p ) B 0.25 0.10 0.20 0.15 0.25 0.23 0.15 0.31
Calvo wages (η w ) B 0.25 0.10 0.27 0.14 0.39 0.32 0.18 0.45
Edgeworth comp. (νnorm ) N 0.50 0.25 0.50 0.47 0.53 0.38 0.27 0.50
Edgeworth comp. (νwar ) N 0.50 0.25 0.35 0.30 0.40 − − −
War draft (p` ) B 0.15 0.05 0.08 0.02 0.14 − − −
War default (pb ) B 0.15 0.05 0.08 0.03 0.13 − − −
War capital depreciation (pδ ) B 0.15 0.05 0.09 0.02 0.15 − − −
Investment adj. cost (φi ) N 5.00 0.50 4.04 3.21 4.89 4.50 3.76 5.25
Trade elasticity (µ) N 1.50 0.25 1.48 1.31 1.65 1.53 1.36 1.68
Labor tax rule (dτ n ) IG 0.25 0.25 0.09 0.07 0.12 0.11 0.07 0.14
Capital tax rule (dτ n ) IG 0.25 0.25 0.13 0.08 0.18 0.14 0.08 0.22
Spending rule (dg ) IG 0.25 0.25 0.15 0.09 0.21 0.11 0.07 0.15
War spillovers
Productivity (φa ) N 0.00 2.00 −0.05 −0.08 −0.02 − − −
Spending (φg ) N 0.00 2.00 0.01 −0.09 0.11 − − −
Investment (φς ) N 0.00 2.00 −0.62 −1.08 −0.17 − − −
Trade (φy∗ ) N 0.00 2.00 −0.70 −0.94 −0.45 − − −
Labor tax (φτ n ) N 0.00 2.00 −0.18 −0.25 −0.10 − − −
Capital tax (φτ n ) N 0.00 2.00 0.38 0.18 0.59 − − −
Money growth (φµ ) N 0.00 2.00 0.03 −0.03 0.10 − − −
Persistence
Productivity (ρa ) B 0.70 0.20 0.98 0.97 1.00 0.99 0.97 1.00
Public spending (ρg ) B 0.70 0.20 0.99 0.98 1.00 1.00 0.99 1.00
Investment (ρζ ) B 0.70 0.20 0.39 0.26 0.52 0.40 0.20 0.59
Foreign demand (ρy∗ ) B 0.70 0.20 0.65 0.58 0.71 0.66 0.59 0.73
Meas. error (ρy
err ) B 0.70 0.20 0.13 0.03 0.22 0.21 0.04 0.37
Labor tax (ρτ n ) B 0.70 0.20 1.00 1.00 1.00 1.00 1.00 1.00
Capital tax (ρτ k ) B 0.70 0.20 0.88 0.82 0.94 0.87 0.82 0.93
Money growth (ρm ) B 0.70 0.20 0.14 0.03 0.24 0.19 0.04 0.33
Standard deviations of shocks

Productivity IG 0.10 Inf 0.06 0.05 0.07 0.08 0.06 0.10
Public spending IG 0.10 Inf 0.10 0.09 0.12 0.06 0.01 0.11
Investment IG 0.10 Inf 0.78 0.59 0.96 0.81 0.63 1.00
Foreign demand IG 0.10 Inf 0.38 0.33 0.44 0.31 0.18 0.45
Meas. error IG 0.10 Inf 0.10 0.09 0.11 0.09 0.07 0.11
Labor tax IG 0.10 Inf 0.06 0.05 0.06 0.05 0.04 0.06
Capital tax IG 0.10 Inf 0.14 0.12 0.16 0.13 0.11 0.15
Money growth IG 0.10 Inf 0.13 0.11 0.16 0.13 0.11 0.16
War IG 0.10 Inf 0.33 0.29 0.37 0.33 0.29 0.37
Marginal data density

1113.42 1102.40
Notes: Results based on 250 000 replications of the MH algorithm. Standard deviations are expressed in percents. N , B and IG respectively
denote Normal, Beta and Inverse Gamma distributions. Marginal data density is the harmonic mean.
50
Table 6: Country-specific estimated parameters for Germany
Dist. Mean Sd. Mean Inf. Sup. Mean Inf. Sup.
Calvo prices (η p ) B 0.25 0.10 0.22 0.16 0.27 0.27 0.20 0.33
Calvo wages (η w ) B 0.25 0.10 0.40 0.33 0.47 0.39 0.31 0.46
War draft (p` ) B 0.15 0.05 0.08 0.02 0.15 − − −
War default (pb ) B 0.15 0.05 0.10 0.02 0.16 − − −
War spillovers
Spending (φg ) N 0.00 2.00 0.03 −0.10 0.17 − − −
Investment (φς ) N 0.00 2.00 −0.46 −0.91 −0.02 − − −
Trade (φy∗ ) N 0.00 2.00 −0.08 −0.29 0.12 − − −
Labor tax (φτ n ) N 0.00 2.00 0.44 0.24 0.63 − − −
Capital tax (φτ n ) N 0.00 2.00 −1.75 −2.49 −1.00 − − −
Money growth (φµ ) N 0.00 2.00 −0.01 −0.07 0.05 − − −
Persistence
Investment (ρζ ) B 0.70 0.20 0.59 0.45 0.73 0.55 0.42 0.68
Meas. error (ρy
err ) B 0.70 0.20 0.24 0.10 0.39 0.24 0.10 0.38
Labor tax (ρτ n ) B 0.70 0.20 0.85 0.75 0.95 0.82 0.74 0.91
Capital tax (ρτ k ) B 0.70 0.20 0.53 0.34 0.71 0.55 0.39 0.73
Money growth (ρm ) B 0.70 0.20 0.38 0.05 0.68 0.67 0.51 0.83

Meas. error IG 0.10 Inf 0.05 0.04 0.06 0.06 0.05 0.07
Labor tax IG 0.10 Inf 0.16 0.13 0.19 0.18 0.16 0.21
War IG 0.10 Inf 0.31 0.27 0.34 0.31 0.27 0.34

1186.84 1177.49
51
Table 7: Country-specific estimated parameters for the United Kingdom
Dist. Mean Sd. Mean Inf Sup Mean Inf Sup
Calvo prices (η p ) B 0.25 0.10 0.25 0.20 0.31 0.33 0.31 0.34
Calvo wages (η w ) B 0.25 0.10 0.14 0.05 0.22 0.38 0.36 0.40
War draft (p` ) B 0.15 0.05 0.11 0.03 0.18 − − −
War default (pb ) B 0.15 0.05 0.06 0.02 0.10 − − −
War spillovers
Spending (φg ) N 0.00 2.00 0.34 0.22 0.47 − − −
Investment (φς ) N 0.00 2.00 −0.34 −0.64 −0.05 − − −
Trade (φy∗ ) N 0.00 2.00 −0.38 −0.52 −0.24 − − −
Labor tax (φτ n ) N 0.00 2.00 −0.14 −0.20 −0.09 − − −
Capital tax (φτ n ) N 0.00 2.00 0.22 0.08 0.36 − − −
Money growth (φµ ) N 0.00 2.00 0.02 −0.00 0.05 − − −
Persistence
Investment (ρζ ) B 0.70 0.20 0.51 0.40 0.62 0.68 0.61 0.72
Meas. error (ρy
err ) B 0.70 0.20 0.15 0.05 0.25 0.29 0.26 0.35
Labor tax (ρτ n ) B 0.70 0.20 1.00 0.99 1.00 1.00 0.99 1.00
Capital tax (ρτ k ) B 0.70 0.20 0.97 0.94 1.00 0.96 0.94 0.99
Money growth (ρm ) B 0.70 0.20 0.25 0.06 0.43 0.46 0.43 0.49

Meas. error IG 0.10 Inf 0.08 0.07 0.09 0.09 0.08 0.10
Labor tax IG 0.10 Inf 0.05 0.04 0.05 0.05 0.04 0.05
War IG 0.10 Inf 0.30 0.27 0.32 0.30 0.27 0.33

1883.54 1848.58
52
Table 8: Country-specific estimated parameters for the United States
Dist. Mean Sd. Mean Inf Sup Mean Inf Sup
Calvo prices (η p ) B 0.25 0.10 0.40 0.36 0.44 0.35 0.34 0.36
Calvo wages (η w ) B 0.25 0.10 0.20 0.11 0.30 0.26 0.26 0.27
War draft (p` ) B 0.15 0.05 0.06 0.02 0.11 − − −
War default (pb ) B 0.15 0.05 0.12 0.06 0.18 − − −
War spillovers
Productivity (φa ) N 0.00 2.00 0.01 −0.01 0.02 − − −
Spending (φg ) N 0.00 2.00 0.14 0.07 0.21 − − −
Investment (φς ) N 0.00 2.00 0.06 −0.31 0.41 − − −
Trade (φy∗ ) N 0.00 2.00 0.16 0.06 0.26 − − −
Labor tax (φτ n ) N 0.00 2.00 0.13 0.06 0.20 − − −
Capital tax (φτ n ) N 0.00 2.00 −0.34 −0.54 −0.14 − − −
Money growth (φµ ) N 0.00 2.00 −0.02 −0.04 0.01 − − −
Persistence
Investment (ρζ ) B 0.70 0.20 0.55 0.43 0.67 0.46 0.44 0.48
Meas. error (ρy
err ) B 0.70 0.20 0.15 0.04 0.24 0.50 0.49 0.52
Labor tax (ρτ n ) B 0.70 0.20 1.00 1.00 1.00 1.00 0.99 1.00
Capital tax (ρτ k ) B 0.70 0.20 0.81 0.75 0.88 0.85 0.84 0.86
Money growth (ρm ) B 0.70 0.20 0.41 0.23 0.59 0.65 0.64 0.67

Meas. error IG 0.10 Inf 0.05 0.04 0.05 0.05 0.05 0.06
Labor tax IG 0.10 Inf 0.08 0.07 0.08 0.08 0.07 0.10
War IG 0.10 Inf 0.29 0.26 0.32 0.28 0.26 0.30

1870.81 1816.41
53
D Prior and posterior distributions
Figure 10: Priors and Posteriors - France
SE_ea SE_eg SE_eys etap etaw nunorm

40 15
80
20
30 10 4
60 10
40 20
5 2 10
5
20 10
0 0 0 0 0 0
0.1 0.2 0.3 0.4 0.5 0.1 0.2 0.3 0.4 0.5 0.2 0.4 0.6 0 0.2 0.4 0.6 0 0.2 0.4 0.6 0.4 0.6 0.8 1 1.2 1.4
SE_ei SE_err SE_etn nuwar pell pdef

15
80 10
10 10
10 40
60
40 5
20 5 5
5
20
0 0 0 0 0 0
0 0.5 1 1.5 0.1 0.2 0.3 0.4 0.5 0.1 0.2 0.3 0.4 0.5 0.2 0.4 0.6 0.8 1 1.2 1.4 0 0.1 0.2 0.3 0 0.1 0.2 0.3
SE_etk SE_em SE_ewar pdelt phii mu

40 30 10 0.8 4
15 3
30 0.6
20
20 10 5 0.4 2
10
10 5 0.2 1
0 0 0 0 0 0
0.1 0.2 0.3 0.4 0.5 0.1 0.2 0.3 0.4 0.5 0.1 0.2 0.3 0.4 0.5 0 0.2 0.4 2 4 6 1 1.5 2
dtn dtk dg phiwm rhoa rhog
10 10 60
40
20
10
30 40
5 5 20
10 5 20
10
0 0 0 0 0 0
0 0.5 1 0 0.5 1 0 0.5 1 −5 0 5 0.2 0.4 0.6 0.8 1 0.2 0.4 0.6 0.8 1
phiwa phiwg phiwi rhoi rhoys rhoerr
20 1.5 10 6
6
4
15
4 1 4
10 5
2
2 0.5 2
5
0 0 0 0 0 0
−5 0 5 −5 0 5 −5 0 5 0 0.2 0.4 0.6 0.8 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8
phiwys phiwtn phiwtk rhotn rhotk rhom
10 6
8 3
400
2 6 4
2
5
4 200
1 2
1
2
0 0 0 0 0 0
−5 0 5 −5 0 5 −5 0 5 0.2 0.4 0.6 0.8 1 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8
54
Figure 11: Priors and Posteriors - Germany

20
80 40 10 10
15
60 15
30
10
40 20 10 5 5
5 5
20 10
0 0 0 0 0 0
0.1 0.2 0.3 0.4 0.5 0.1 0.2 0.3 0.4 0.5 0.1 0.2 0.3 0.4 0.5 0 0.2 0.4 0.6 0.2 0.4 0.6 0.4 0.6 0.8 1 1.2 1.4

10
15 20 10
80 8
60 15
10 6
5 5
40 10 4
5
20 5 2
0 0 0 0 0 0
0.2 0.4 0.6 0.8 1 0.1 0.2 0.3 0.4 0.5 0.1 0.2 0.3 0.4 0.5 0.2 0.4 0.6 0.8 1 1.2 1.4 0 0.2 0.4 0 0.2 0.4
15 15 0.8 4
20 8
6 0.6 3
10 10 15
4 0.4 2
10
5 5 0.2 1
5 2
0 0 0 0 0 0
0.1 0.2 0.3 0.4 0.5 0 0.2 0.4 0.1 0.2 0.3 0.4 0.5 0 0.2 0.4 2 4 6 8 1 1.5 2

6 40
10 20 80
30
4 15 10 60
5 40 20
10
2 5
5 20 10
0 0 0 0 0 0
0 0.5 1 0 0.5 1 0 0.5 1 −5 0 5 0.2 0.4 0.6 0.8 1 0.2 0.4 0.6 0.8 1
20 1.5 4
4 4
4
15 1 3 3
10 2 2 2 2
0.5
5 1 1
0 0 0 0 0 0
−5 0 5 −5 0 5 −5 0 5 0.2 0.4 0.6 0.8 1 1.2 0 0.5 1 −0.2 0 0.2 0.4 0.6 0.8
6 2
3 3 0.8 3
0.6 1.5
2 2 4 2
0.4 1
1 1 2 1
0.2 0.5
0 0 0 0 0 0
−5 0 5 −5 0 5 −5 0 5 0.2 0.4 0.6 0.8 1 0 0.5 1 0 0.5 1
55
Figure 12: Priors and Posteriors - UK

8
40
150
40 20 10 6 30
100
4 20
20 10 5
50 2 10
0 0 0 0 0 0
0.1 0.2 0.3 0.4 0.5 0.1 0.2 0.3 0.4 0.5 0.1 0.2 0.3 0.4 0.5 0.2 0.4 0.6 0 0.2 0.4 0.6 0.4 0.6 0.8 1 1.2 1.4

80 15 10
15
100 15
60
10 10
10
40 5
50
5 5 5
20
0 0 0 0 0 0
0.2 0.4 0.6 0.8 1 0.1 0.2 0.3 0.4 0.5 0.1 0.2 0.3 0.4 0.5 0.4 0.6 0.8 1 1.2 1.4 0 0.2 0.4 0 0.1 0.2 0.3

60 10 0.8 8
40 20
0.6 6
40 30 15
5 0.4 4
20 10
20
10 5 0.2 2
0 0 0 0 0 0
0.1 0.2 0.3 0.4 0.5 0 0.2 0.4 0.1 0.2 0.3 0.4 0.5 0 0.2 0.4 2 4 6 1 1.5 2

40 20
40
15 15
30 20 15
30
10 10
20 10 20
10
10 5 5 5 10
0 0 0 0 0 0
0.2 0.4 0.6 0.8 1 0 0.5 1 0 0.5 1 −5 0 5 0.2 0.4 0.6 0.8 1 0.2 0.4 0.6 0.8 1

40 6 6
2 8
30 4 6
1.5 4 4
20 4
1
2 2 2
10 0.5 2
0 0 0 0 0 0
−5 0 5 −5 0 5 −5 0 5 0.2 0.4 0.6 0.8 1 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8

20
150 3
4 10 4 15
100 2
10
2 5 2
50 5 1
0 0 0 0 0 0
−5 0 5 −5 0 5 −5 0 5 0.2 0.4 0.6 0.8 1 0.2 0.4 0.6 0.8 1 0 0.5 1
56
Figure 13: Priors and Posteriors - US

200 20 40
80 15 6
150 15 30
60
10 4
100 40 10 20
50 5 5 2 10
20
0 0 0 0 0 0
0.1 0.2 0.3 0.4 0.5 0.1 0.2 0.3 0.4 0.5 0.1 0.2 0.3 0.4 0.5 0.2 0.4 0.6 0 0.2 0.4 0.6 0.4 0.6 0.8 1 1.2 1.4

15 15
60 15 10
100 10
10
40 10
50 5
5 5
20 5
0 0 0 0 0 0
0.2 0.4 0.6 0.8 1 1.2 0.1 0.2 0.3 0.4 0.5 0.1 0.2 0.3 0.4 0.5 0.4 0.6 0.8 1 1.2 1.4 0 0.1 0.2 0.3 0 0.1 0.2 0.3

10 8
0.8
30 20
40 6
15 0.6
20 5 4
10 0.4
20
10 0.2 2
5
0 0 0 0 0 0
0.1 0.2 0.3 0.4 0.5 0 0.2 0.4 0.1 0.2 0.3 0.4 0.5 0 0.2 0.4 2 4 6 8 1 1.5 2

15
20 400
8
10 6 15 20 20 300
4 10 200
5 10 10
2 5 100
0 0 0 0 0 0
0 0.5 1 0 0.5 1 0 0.5 1 −5 0 5 0.2 0.4 0.6 0.8 1 0.2 0.4 0.6 0.8 1

2 6
10 6
30 1.5 10
4
4
20 5 1
2 5
2
10 0.5
0 0 0 0 0 0
−5 0 5 −5 0 5 −5 0 5 0.2 0.4 0.6 0.8 1 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8

10 10
6 3 300
3
2 200 2
4 5
5
100 1
2 1
0 0 0
0 0 0 0.2 0.4 0.6 0.8 1 0.2 0.4 0.6 0.8 1 0 0.5 1
−5 0 5 −5 0 5 −5 0 5
57
E Smoothed innovations to the structural shocks
Figure 14: Smoothed innovations to the structural shocks - France

Productivity Public spending
0.2 0.3
0.2
0.1
0.1
0 0
−0.1
−0.1
−0.2
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000
Foreign demand Money

1 0.4
0.5
0.2
0
0
−0.5
−0.2
−1
−1.5 −0.4
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000
Investment efficiency Labor tax

0.2
1
0.1
0
0
−1
−0.1
−2
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000
Capital tax War

0.4 1
0.2 0.8
0 0.6
−0.2 0.4
−0.4 0.2
−0.6 0
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000
58
Figure 15: Smoothed innovations to the structural shocks - Germany
0.2
0.1
0
0.05
0 −0.2
−0.05 −0.4
−0.1 −0.6
1900 1920 1940 1960 1980 2000 1900 1920 1940 1960 1980 2000

0.4 0.4
0.2 0.3
0
0.2
−0.2
0.1
−0.4
0
−0.6
−0.1
1900 1920 1940 1960 1980 2000 1900 1920 1940 1960 1980 2000

1 1
0.5
0.5
0
−0.5 0
−1
−0.5
−1.5
1900 1920 1940 1960 1980 2000 1900 1920 1940 1960 1980 2000
Capital tax War

1
0.5 0.8
0.6
0 0.4
0.2
−0.5
0
1900 1920 1940 1960 1980 2000 1900 1920 1940 1960 1980 2000
59
Figure 16: Smoothed innovations to the structural shocks - United Kingdom
0.1
0.6
0.05
0.4
0 0.2
−0.05 0
−0.2
1880 1900 1920 1940 1960 1980 2000 1880 1900 1920 1940 1960 1980 2000
0.5 0.1
0
0
−0.1
−0.5
−0.2
1880 1900 1920 1940 1960 1980 2000 1880 1900 1920 1940 1960 1980 2000

1.5
0.1
1
0.5 0
0
−0.1
−0.5
−1
−0.2
1880 1900 1920 1940 1960 1980 2000 1880 1900 1920 1940 1960 1980 2000
Capital tax War

1
0.2 0.8
0.6
0
0.4
−0.2 0.2
0
1880 1900 1920 1940 1960 1980 2000 1880 1900 1920 1940 1960 1980 2000
60
Figure 17: Smoothed innovations to the structural shocks - United States
0.4
0.05 0.3
0.2
0 0.1
0
−0.05 −0.1
1880 1900 1920 1940 1960 1980 2000 1880 1900 1920 1940 1960 1980 2000

0.1
0.5
0.05
0
0
−0.05
−0.1
−0.5
1880 1900 1920 1940 1960 1980 2000 1880 1900 1920 1940 1960 1980 2000

2 0.3
1 0.2
0.1
0
0
−1 −0.1
−2 −0.2
1880 1900 1920 1940 1960 1980 2000 1880 1900 1920 1940 1960 1980 2000
Capital tax War

1
0.5
0.8
0 0.6
0.4
−0.5
0.2
−1 0
1880 1900 1920 1940 1960 1980 2000 1880 1900 1920 1940 1960 1980 2000
61
F The macroeconomic effects of a shock on the war indicator
variable in the US
Based on 120 simulations, Figures 18 and 19 report the effects of a 5-years war episode, along with the
associated confidence intervals for macro variables and for policy variables respectively.
Figure 18: Impulse Response Functions to a 5-year shock on

∆t - Macro variables
Shaded areas correspond to one standard deviation confidence intervals

(16th and 84th percentiles) based on 120 simulations.
First, the capital stocks (both private and public) depreciate, which contributes negatively on impact to
the dynamics of output but boosts public and private investment with positive impact on output in the
subsequent periods. The effect on consumption is clearly negative as savings must increase to finance the
rise in investment. Second, a war-related partial default occurs, which, all else equal, tends to lower the
debt-to-GDP ratio and lowers tax rates through the fiscal rules, with positive expected effects on output
dynamics. As shown in the next set of figures, it happens that these two effects play a relatively minor role
from a quantitative perspective and are therefore not at the center of the stage. Third, a military draft
lowers the bulk of hours worked used in the production process, which tends to put upward pressure on
equilibrium wages, lowers output and consumption. This channel has quantitatively limited implications
as well, although to a lesser extent than capital depreciation and partial default.
Spillover terms play a much bigger role. The shock on the war indicator variable is associated with a
positive foreign demand shock that should lead to an improvement of the trade balance. However, this
effect is more than compensated by the public spending spillover. Its importance is massive, as depicted
by Figure 20 and 21. The rise of public spending is very large – public expenditure almost double –
and drives the positive dynamics of output, the crowding-out effect on investment, the negative (positive)
62
Figure 19: Impulse Response Functions to a 5-year shock on
∆t - Policy variables
Shaded areas correspond to one standard deviation confidence intervals

(16th and 84th percentiles) based on 120 simulations.
effect on consumption (hours worked), and the terms-of-trade appreciation. It generates a trade deficit
in the first periods as well as a massive increase in the debt-to-GDP ratio. Finally, tax spillovers imply
a large rise in the labor income tax and a fall in the capital income tax, as was expected given the
signs of estimated spillover coefficients (see Table 8 in Appendix C). The labor tax tends to cut GDP
and consumption while the capital tax should act as a subsidy to investment and should boost output.
However, as already mentioned, the overall dynamics of the economy seem almost entirely driven by the
rise in public spending.
How does our analysis compare with existing papers? Quite well actually, in the sense that most of
them find that the dynamics of the US economy during WWII are well captured by a large rise in public
spending (see McGrattan and Ohanian (2010)). In particular, the positive joint dynamics of hours worked,
pre-tax wages, GDP and the negative dynamics of investment identified in the data by McGrattan and
Ohanian (2010) are correctly reproduced, at least qualitatively. Similarly, our model predicts a rise in the
labor income tax and a rise in public debt, as observed in US data. Most of the additional channels of our
model (military draft, partial default, capital depreciation) contribute only marginally to the dynamics of
the economy and are quantitatively almost negligible compared to the effects of the massive rise in public
spending.
G Historical decompositions for France, Germany and the UK
63
Figure 20: IRFs to a 5-year shock on ∆t - Sensitivity - Macro variables
GDP Consumption Investment
0 0
5
−5
0
% dev.
% dev.
% dev.
−5 −10
−5 −15
−20
−10 −10
−25
5 10 15 20 25 5 10 15 20 25 5 10 15 20 25
Capital stock Hours Real wage
0 10
5
5
−5
% dev.
% dev.
% dev.
0 0
−10 −5
−5
−10
−15
5 10 15 20 25 5 10 15 20 25 5 10 15 20 25
Net exports to GDP Terms of trade Inflation
0
2 −5 5
% dev.
% dev.
−10
1 0
%
−15
0
−20 −5
−1 −25
5 10 15 20 25 5 10 15 20 25 5 10 15 20 25
Years Years Years
Black: baseline model, dashed: no spillovers, dotted: no draft
Figure 21: IRFs to a 5-year shock on ∆t - Sensitivity - Policy variables

Debt-to-GDP Real balances
0
90
−5
80
% dev.
%
70 −10
60 −15
50
5 10 15 20 25 5 10 15 20 25
Public invt Public consumption
80
4
60
% dev.
% dev.
40
2 20
0
0 −20
5 10 15 20 25 5 10 15 20 25
Labor tax Capital tax
35 40
30 30
Level
Level
25 20
10
20
0
5 10 15 20 25 5 10 15 20 25
Years Years
Black: baseline model, dashed: no spillovers, dotted: no draft
64
Figure 22: France - historical decomposition

0.6 0.6
War War
0.4 0.4
Money growth Money growth
0.2 0.2
0 0
-0.2 -0.2
-0.4 -0.4
Taxes Taxes
-0.6 -0.6
0 5 10 15 20 25 30 35 40 45 50 0 5 10 15 20 25 30 35 40 45 50
65
0.1 0.15
War 0.1 War
0.05
0.05
TFP + investment -0.05 TFP + investment
-0.1
-0.05
-0.15
Foreign demand -0.2 Foreign demand

-0.1
-0.25
Taxes Taxes
-0.15 -0.3
0 5 10 15 20 25 30 35 40 45 50 0 5 10 15 20 25 30 35 40 45 50
Figure 23: Germany - historical decomposition

0.4 1.4
War 1.2 War

0.2
1
0
0.8
-0.2
TFP + investment 0.6 TFP + investment
-0.4
0.4
-0.6 Public spending Public spending
0.2
-0.8
Foreign demand 0 Foreign demand
-1
-0.2
Taxes Taxes
-1.2 -0.4
0 5 10 15 20 25 30 35 40 45 50 0 5 10 15 20 25 30 35 40 45 50
66
0.15 0.08
War 0.06 War

0.1
0.04
0.05 Money growth Money growth
0.02
0
-0.05
Public spending -0.02 Public spending
-0.1
-0.04
-0.15
-0.06
Taxes Taxes
-0.2 -0.08
0 5 10 15 20 25 30 35 40 45 50 0 5 10 15 20 25 30 35 40 45 50
Figure 24: UK - historical decomposition

0.5 0.4
0.4 War 0.3 War
0.3 0.2
0.2 0.1
0.1 0
0 Public spending -0.1 Public spending
-0.1 -0.2
-0.2 -0.3
Taxes Taxes
-0.3 -0.4
0 5 10 15 20 25 30 35 40 45 50 0 5 10 15 20 25 30 35 40 45 50
67
0.15 0.08
0.06
War War
0.1
0.04

0.02
0.05
0 -0.02
-0.04
-0.05
-0.06

-0.08
-0.1
-0.1
Taxes Taxes
-0.15 -0.12
0 5 10 15 20 25 30 35 40 45 50 0 5 10 15 20 25 30 35 40 45 50
H The hyperinflation in Germany and the Great Depression in
the US
What does our model predict for those two particular periods? More precisely, what are the shocks that
account for the data around these dates according to our estimated model? Figure 25 below shows the
model innovations to the structural shocks produced by our estimation of the model for Germany and the
US during these historical episodes.
Figure 25: Smoothed innovations to the structural shocks
(a) Germany (b) US

Productivity Public spending Productivity Public spending
0.2
0.1 0.05 0.1
0.1
0.05
0
0 0.05
0
−0.1
−0.05
−0.2
−0.05 0
−0.1
−0.3
1916 1917 1918 1919 1920 1921 1922 1923 1924 1916 1917 1918 1919 1920 1921 1922 1923 1924 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938
Foreign demand Money Foreign demand Money

0.4 0.2 0
0.1
0.3 0.1
0.2 0 −0.05
0
0.1 −0.1
−0.1
−0.1 0 −0.2
−0.3
−0.1
1916 1917 1918 1919 1920 1921 1922 1923 1924 1916 1917 1918 1919 1920 1921 1922 1923 1924 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938
Investment efficiency Labor tax Investment efficiency Labor tax

0.8 0.5 0.15
0.6 1 0
0.1
0.4 −0.5
0.05
0.2 0.5 −1
0 −1.5 0
−0.2 −2 −0.05
0
1916 1917 1918 1919 1920 1921 1922 1923 1924 1916 1917 1918 1919 1920 1921 1922 1923 1924 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938
Capital tax War Capital tax War

1 1
0.4
0.8 0 0.5
0.2 0.6
−0.1 0
0 0.4
−0.2 −0.5
−0.2 0.2
0 −1
1916 1917 1918 1919 1920 1921 1922 1923 1924 1916 1917 1918 1919 1920 1921 1922 1923 1924 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938
Starting with Germany, Figure 25a shows that the hyperinflation of the early 20’s episode is captured
by our model through a series of large positive shocks to the money growth rate, along with a peak in
the labor income tax. The latter probably helps capturing the large fall in hours worked and the rise
in unemployment observed in Germany at that time. In addition, a moderate increase in investment
efficiency appears as the last “driver” of the economy over this period.
Now let us look at smoothed shocks during the Great Depression, reported in Figure 25b. The fall in
GDP in the early years of the Great Depression is captured by our model through a negative shock on
investment efficiency. This is consistent with the fact that the Great Depression was arguably generated
by a crash of financial markets and resulted in a twisted channeling of savings to investment opportunities.
In addition, the resulting trade collapse is captured by a negative shock on foreign demand. Deflationary
pressures are captured by negative shocks to the money growth rate, starting in 1929 and returning to
normal in 1934. The model also finds a large increase in the labor income tax. Here again, we believe
that this is the way our model copes with the large rise in unemployment. Last but not least, our model
predicts a small discretionary rise in public spending in 1934, and a much larger rise in 1936, certainly
reflecting the New Deal.
68

Episodes of War and Peace in An Estimated Open Economy Model

Uploaded by

Copyright:

Available Formats

Episodes of War and Peace in An Estimated Open Economy Model

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Episodes of War and Peace in An Estimated Open Economy Model

Uploaded by

Copyright:

Available Formats

Version of Record: https://www.sciencedirect.

Episodes of War and Peace in an

June 13, 2019

JEL Classification: E31, E32, E62, F41, H56, H63.

In the welfare index, total consumption e

νt = (1 − ∆t ) νnorm + ∆t νwar (3)

In addition, capital accumulation is subject to investment adjustment costs and to investment

at = (1 − ρa ) + ρa at−1 + φa ∆t + εat (7)

gt = g + dτn (brt−1 − br) + φg ∆t + sgt (10)

sgt = ρg sgt−1 + εgt (13)

We consider that money supply grows at a rate µm

dy obst = log(yt ) − log(yt−1 ) + yterr (18)

Xt = A (Θs , Θc ) Xt−1 + B (Θs , Θc ) εt , where εt ∼ N (0, Σ) (20)

Mt = C (Θs , Θc ) + D (Θs , Θc ) Xt (22)

P (Θs |MT ) ∝ L (MT |Θs ) P (Θs ) (24)

Country-specific calibrated parameters. Remaining calibrated parameters are country-specific, to

Table 2: Calibrated country-specific parameter values (Θc2 )

5 United States: estimation results and implications

We rather focus on the Edgeworth complementarity parameter ν, and war-specific parameters

5.2 Historical decompositions

(a) GDP growth (b) Inflation rate

0.2 Money growth Money growth

Foreign demand Foreign demand

0 Money growth 0.04 Money growth

Public spending Public spending

Foreign demand Foreign demand

5.3 Public spending multipliers

denote the multiplier in normal times and

denote the multiplier conditional on a 5-years war episode.

crowding-out of private consumption.

Table 3: Key parameter mean estimates, with 90% confidence intervals

Notes: Results are extracted from Tables 5 to 8 in Appendix C.

6.3 The macroeconomic effects of a simulated war episode

Figure 4: Impulse Response Functions to a 5-year MIT shock on ∆t - Policy variables

6.4 Public spending output multipliers

(a) France (b) Germany

7 Welfare losses from war episodes

Anderton, C. H. and J. R. Carter, The Impact of War on Trade: An Interrupted Times-Series

Auerbach, A. J. and Y. Gorodnichenko, Fiscal Multipliers in Recession and Expansion, in Fiscal

, , and P. Gomme, A Tale of Tax Policies in Open Economies, International Economic

Barro, R. J. and J. F. Ursúa, Stock-market Crashes and Depressions, Research in Economics,

Devereux, M. B. and G. W. Smith, Transfer Problem Dynamics: Macroeconomics of the Franco-

Furlanetto, F. and M. Seneca, New Perspectives on Capital Depreciation Shocks as Sources of

Lucas, R. E., Models of business cycles, Oxford: Blackwell, 1987.

1900 1950 2000 1900 1950 2000 1900 1950 2000

1900 1950 2000 1900 1950 2000 1900 1950 2000

1900 1950 2000 1900 1950 2000 1900 1950 2000

0.15 0.04 0.1

In the welfare index, total consumption e

The budget constraint of agent j is

+ (1 − τtn ) wt (j) nt (j) + Πt (j) + taxt (j) (B.4)

ζt = (1 − ρζ ) + ρζ ζt−1 + φζ ∆t + εζt (B.7)

at = (1 − ρa ) + ρa at−1 + φa ∆t + εat (B.22)

Cost minimization implies

taking into account the demand addressed to firm ω

The optimal pricing condition is thus

while the dynamics of prices are given by

gt = g + dτn (brt−1 − br) + φg ∆t + sgt (B.30)

sgt = ρg sgt−1 + εgt (B.33)