The Price of War

DSGE model documentation

Jonathan Federle, André Meier, Gernot J. Müller,

Willi Mutschler, Moritz Schularick
May 2024

NOT FOR PUBLICATION

This document derives in detail the multi-country DSGE model used in section 4 of the
paper The Price of War. Replication codes are written in Dynare and available at

[Link]

Please raise any issues or questions as an issue on GitHub (or get in touch via email).

1
Contents

1 Overview 3

2 Notation 5

3 Households 7
3.1 Utility function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.2 Capital accumulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.3 Budget constraint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.4 Stochastic discount factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.5 Utility maximization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.6 Labor supply and wage setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.6.1 Labor aggregation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.6.2 Wage setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

4 Firms 19
4.1 Retail goods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
4.2 Wholesale goods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.2.1 Export good aggregation (PCP) . . . . . . . . . . . . . . . . . . . . . . . . . . 24
4.2.2 Local good aggregation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.3 Intermediate goods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.3.1 Production function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.3.2 Profit maximization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.3.3 Marginal costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.3.4 Factor demands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.3.5 Export pricing (PCP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
4.3.6 Local pricing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

5 Government 38
5.1 Monetary policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
5.2 Fiscal policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

6 Market clearing 40
6.1 Capital market . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
6.2 Labor market . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
6.3 Aggregation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
6.4 Bond market . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
6.5 Aggregate demand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
6.6 Aggregate supply . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
6.7 Aggregate budget constraint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

7 War shock scenario 47

2
1 Overview
• Dynamic Stochastic General Equilibrium (DSGE) model with multiple countries denoted by
Home (H), Nearby (N ), Distant (D), and Rest of World (R).
• Focus is on the impact of a war shock in Home and the international spillovers.

Households

• Households maximize lifetime utility considering consumption, labor supply and habit formation.
• Capital accumulation is governed by a law of motion with investment adjustment costs.
• Monopolistic competitive households set wage for distinct labor types according to Calvo frictions.
• Households hold foreign bonds and use imported goods for domestic absorption.
• Budget constraint includes income from labor, capital, bonds and government transfers.

Firms

• Retail goods are distributed using a CES aggregation of locally produced and imported goods,
subject to import adjustment costs and iceberg trade frictions.
• Wholesale goods are produced by monopolistic competitive firms using labor, capital and inter-
mediate inputs in production. Price setting is subject to Calvo price frictions.
• Export prices are set under the producer currency pricing (PCP) regime.

Government

• Monetary Policy is defined by an interest rate rule with feedback to inflation deviations and
output gap.
• Fiscal authority finances government spending by issuing bonds and lump-sum taxes.

Market clearing

• Sequence-of-markets concept ensures equilibrium in capital, labor, bond and goods markets.
• Price and wage inefficiencies are captured by dispersion terms.
• Test of market clearing by including all aggregated budget constraints (Walras law).

War shock

• Autoregressive process of order two.

• Home’s productivity declines, trade costs with Home go up and capital is destroyed in Home.
• All countries increase government spending to finance military expenditures, with varying mag-
nitudes.
• Section 7 provides details which equations are affected by the war shock.

3
Implementation in Dynare

• Model equations are derived in per-capita variables.

• Gray boxes contain model equations as entered in Dynare.

Related literature and references The model and its derivation are inspired by the literature on
open-economy DSGE models and we draw heavily on codes and technical appendices of the following
references:

• Cwik, Müller, and Wolters (2011)

• Eichenbaum, Johannsen, and Rebelo (2021)

• Erceg, Guerrieri, and Gust (2006)

• Georgiadis and Schumann (2021)

• Gomes, Jacquinot, and Pisani (2012)

• Gopinath et al. (2020)

4
2 Notation

The following conventions provide structure to the notation to effectively navigate the international
dimensions and interactions between countries and agents within the model.

Indices

• subscript t is an index for time

• subscript j is a specific index for identifying a single country
• subscript i is an index that loops over all countries
• variables with a double index follow the convention that the first index denotes the origin and
the second index the destination, e.g. Yij,t represents goods produced in country i (origin) and
sold to country j (destination)
• superscripts s and d indicate, respectively, supply or demand side variables, e.g. labor supply
Lsj,t rhs of household h or demand of domestic firm f for locally produced wholesale goods Yjj,t
d rfs

Economic sizes and entities Within the global economy, there exists a continuum of firms and
households, each of whose sizes is normalized to unity. Nj then specifies the portion on the unit
interval attributed to country j, such that nj “ |Nj | quantifies the economic size of both firms and
households of j.

To navigate through the various economic entities the following indices are assigned:

• households are indexed by h P Nj

• labor unions are indexed by ℏ P Nj
• firms in the final goods sector are indexed by f P Nj
• firms in the imported wholesale goods sector are indexed by f P Nj
• firms in the local wholesale goods sector are indexed by f P Nj
• monopolists in the export branch of the intermediate goods sector are indexed by m P Nj
• monopolists in the local branch of the intermediate goods sector are indexed by m P Nj

Square brackets are exclusively used to index decisions by specific economic entities, e.g. Yij,t rf s.
Lowercase letters denote per-capita variables, e.g. yij,t :“ n1j Nj Yij,t rf s df .
ş

Interest rates, wages and prices Uppercase letter Rj,t

K specifies the nominal rental rate of capital in
K
Rj,t
country j at time t. Lower-case letter rj,t
K :“
Pj,t represents the real return as it is deflated by the
consumer price index (CPI) Pj,t .

Uppercase letters such as Wj,t or W

Ăj,t rhs specify nominal wages in country j denominated in the
Wj,t W
Ăj,t rhs
currency of j at time t. Lower-case letters for wages, such as wj,t :“ Pj,t or w
rj,t rhs :“ Pj,t ,
represent real wages as they are deflated by the consumer price index (CPI) Pj,t .

Uppercase letter Pij,t specifies the price of goods produced in country i (origin) and sold to country
j (destination). Prices are generally denominated in the currency of the importing country j, except
for export prices Pij,t , which are denominated in a different currency depending on the currency
P CP,EX

5
pricing paradigm. Note that the subindices do not switch between import and export prices, but
Pij,t
stick to the origin-destination convention. Lower-case letters for prices, such as pij,t :“ Pj,t , represent
relative prices that are deflated by the consumer price index (CPI) Pj,t . Export prices are special as
they need to be deflated by the currency in which they are denominated:
P CP,EX
P CP,EX Pij,t
pij,t :“
Pi,t

Exchange rates The nominal bilateral exchange rate is denoted by Eij,t

n , converting the currency of
P
country i to the currency of country j. The real exchange rate is defined by Eij,t ij,t Pj,t .
r “ En i,t

Bonds Uppercase letter Bij,t rhs specifies the quantity of bonds issued by country i and held by
household h in country j at time t. Bonds are denominated in the currency of the originating country
i. Lowercase letters for bonds represent real per-capita amounts deflated by the consumer price index:
ż
1 Bij,t rhs
bij,t :“ dh
nj Nj Pi,t

Bjj,t rhs Bj,t rhs

The same notation holds for domestic bonds: bjj,t :“ 1
and bj,t :“ 1
ş ş
nj Nj Pj,t dh nj Nj Pj,t dh.

6
3 Households

3.1 Utility function

Life-time utility Household h P Nj seeks to maximize their expected lifetime utility represented as

8
ÿ
Et pβj qk µj,t`k Uj,t`k rhs
k“0

Here, βj is the discount factor, while µj,t represents a time-varying shock to the household’s time
preference which evolves according to an exogenous AR(1) process.

Discount Factor Shifter

discount factor shifter in j

logpµj,t q “ ρµj logpµj,t´1 q ` εµj,t

Felicity function The contemporaneous utility function is given by

1 ` C
˘1´σjC L 1 ` s ˘1`σjL
Uj,t rhs “ C j,t rhs ´ ϕ j cj,t´1 ´ χ j Lj,t rhs
1 ´ σjC 1 ` σjL

Here, σjC is the inverse of the intertemporal elasticity of substitution in consumption Cj,t rhs, σjL is
the inverse of the Frisch elasticity of work effort Lsj,t rhs and χL
j is a weight parameter that scales
labor supply in steady-state. Additionally, the utility from consumption is subject to habit formation,
which is parametrized by ϕC
j . The habit stock is defined as the previous period’s per-capita private
consumption cj,t´1 :“ 1{nj Nj Cj,t´1 rhsdh. This stock is assumed to be external, implying that the
ş

household does not consider its own contribution to cj,t´1 when formulating optimal plans.

3.2 Capital accumulation

Law of motion Each household owns a share Kj,t rhs of the total capital stock in country j, which
evolves according to the following law of motion:

Kj,t rhs “ p1 ´ δjK qKj,t´1 rhs ` ζj,t 1 ´ ΦK

` ˘
j,t rhs Ij,t rhs

where δjK is the depreciation rate of capital and Ij,t rhs is the amount of final retail goods invested in
new capital. ζj,t is a time-varying shock to the efficiency of investment, which is assumed to follow an
AR(1) process.

Investment-Specific Productivity

investment-specific productivity in j

logpζj,t q “ ρζj logpζj,t´1 q ` εζj,t

7
 j,t are convex investment adjustment costs parameterized by ϕj :
Investment adjustment costs ΦK K

ϕK
ˆ ˙2
j Ij,t rhs
ΦK
j,t rhs“ ´1
2 Ij,t´1 rhs
BΦK
ˆ ˙ˆ ˙
j,t rhs K Ij,t rhs 1
“ ϕj ´1
BIj,t rhs Ij,t´1 rhs Ij,t´1 rhs
K ˆ ˙˜ ¸
BΦj,t`1 rhs K Ij,t`1 rhs ´1
“ ϕj ´1 Ij,t`1 rhs
BIj,t rhs Ij,t rhs pIj,t rhsq2

3.3 Budget constraint

The nominal per-period budget constraint of a household h P Nj is expressed in currency j as:

Wj,t rhsLsj,t rhs ` R

looooooomooooooon
K
j,t Kj,t´1 rhs
loooooomoooooon
labor income capital income

` p1 ` ρB B B
j Pj,t qBj,t´1 rhs ´ Pj,t Bj,t rhs ` P
looooooooooooooooooooomooooooooooooooooooooon j,t T Rj,t rhs
loooooomoooooon
domestic government bonds transfers
ż ż
1 local 1 ÿ P CP,EX
` DIV j,t rmsdm ` DIV ji,t rmsdm
n j Nj nj i‰j Nj
loooooooooooooooooooooooooooooooooomoooooooooooooooooooooooooooooooooon
dividend income monopolists
ÿ
B 1
Pj,t Cj,t rhs ` Pj,t Ij,t rhs `
“ looooooooooooomooooooooooooon Pj,t rs sBj,t rs1 , hs ´ Bj,t´1 rs, hs
s1 PS
retail good expenditures looooooooooooooooooooomooooooooooooooooooooon
domestic state-contingent bonds
ÿ ÿ
n
` Eij,t pBij,t rhs ´ Rij,t´1 Bij,t´1 rhsq ` Pj,t ΦB
ij,t rhs
i i‰j
loooooooooooooooooooooomoooooooooooooooooooooon looooooomooooooon
international non-state-contingent bonds international fees

On the income side, every household determines its own salary Wj,t rhs in a monopolistically competitive
labor market. They supply an individual variety of labor Lsj,t rhs, which is aggregated by labor unions
and then offered to the intermediate good producing firms. At the same time, the household rents
out capital Kj,t´1 rhs to this sector at nominal rate Rj,t
K . Households either receive lump-sum transfers

(T Rj,t rhs ą 0) or make lump-sum tax payments (T Rj,t rhs ă 0) to the fiscal authority. They also obtain
P CP,EX
income from dividends, symbolized by DIV ij,t rms and DIV j,t rms, reflecting the household’s stake
local

in firm ownership that is evenly distributed across the population. On the expenditure front, Cj,t rhs
and Ij,t rhs represent the purchases of final retail goods for consumption and investment purposes, each
priced at the consumer price index Pj,t .

In the financial realm, households have access to different types of risk-free securities, each serving
distinct functions within the portfolio.

State-contingent bonds For hedging against individual income risk, households have access to a full
set of domestic state-contingent zero-coupon bonds, in currency j and restricted to domestic trading.
Specifically, define S as the complete set of future states of the world, then Pj,t
B rs1 s represents the price

at time t of a bond that will pay one unit of domestic currency at time t ` 1 if the potential future

8
state s1 P S materializes. The quantity of such bonds purchased by household h at time t is denoted
by Bj,t rs1 , hs. Conversely, Bj,t´1 rs, hs indicates the value of the household’s existing claims based on
the realized state s.1

Government bonds Bj,t rhs represents the household’s purchases of government bonds, which are
denominated in the local currency j and only traded domestically to finance fiscal activities. These
bonds are structured as perpetuities with coupon payments that decay at the rate of ρB
j P r0, 1s as in
Woodford (2001). That is, issuing one unit of government bonds in period t commits the government
2
to a coupon payment of one unit of currency in t ` 1, ρB
j in t ` 2, pρj q in t ` 3 and so on. Let
B

new rhs denote the new issuance of such bonds, then the total coupon liability from past issuances
Bj,t
can be summarized as:

0 1 2
Bj,t rhs “ pρB new B new B new new B
j q Bj,t rhs ` pρj q Bj,t´1 rhs ` pρj q Bj,t´2 rhs ` . . . ðñ Bj,t rhs “ Bj,t rhs ´ ρj Bj,t´1 rhs

The appeal of these decaying coupon bonds lies in their simplicity to model maturity structures: only
the total outstanding stock of bonds needs to be tracked, as the coupon payments adjust automatically.
New issuances in period t are traded at a market price denoted by Pj,t B . Consequently, the nominal
´ ¯
value of the portfolio is given by Bj,t´1 rhs ` Pj,t
B B new rhs “ 1 ´ ρB P B B
j,t j j,t j,t´1 rhs ` Pj,t Bj,t rhs. Under
B

´1
constant prices, the average duration is equal to p1 ´ βj ρB
j q . This implies that when ρB
j “ 0 the
portfolio comprises one-period debt, while at ρB
j “ 1 it consists entirely of consol bonds.

International bonds Households engage in the trading of international non-state-contingent bonds,

which are denominated in the currency of the issuing country i (including the domestic country j) and
are priced at one unit of currency. The corresponding gross yield on these bonds, captured by Rij,t ,
is paid out in the next period. The quantity of international bond purchases at time t is denoted by
Bij,t rhs, while Bij,t´1 rhs reflects the nominal value of the household’s existing claims from the previous
period. To ensure stationarity of the economy’s net foreign asset position, holding international bonds
incurs a real carrying cost of Φij,t rhs. ΦB
ij,t rhs is specified as a convex cost function, parametrized by

ij , designed to penalize deviations of real bond holdings from a predetermined target level bij :
ϕB

˙2 ˜ ¸2
ϕB n B
Eij,t ϕB
ˆ
ij ij,t rhs ij
ΦB
ij,t rhs “ ´ bij “ r
loEomo
ij,ton bij,t rhs ´bij
2 Pj,t 2 loomoon
P i,t Bij,t rhs
n
Eij,t P j,t Pi,t

BΦB
ˆ n ˙ n
ij,t rhs Eij,t Bij,t rhs Eij,t
“ ϕB
ij ´ bij
BBij,t rhs Pj,t Pj,t

1
A more elaborate exposition can be achieved by refining the notation to reflect more explicitly the dependency of
variables on the history of states. For example, Cj,t rhs could be written as Cj rst , hs, where st encapsulates the history
of all realized states up to time t. One would then introduce a state-dependent probability πpst`1 |st q to move from
st to st`1 and zoom in on the corresponding expectation operation in the utility maximization problem. Focusing on
the probabilistic framework then effectively discretizes the array of potential future states, allowing expectations to
be honed in on. This approach ultimately has the same result as above, namely that state contingency is implicitly
captured by the asset pricing kernel and one can directly use the streamlined and more compact notation.

9
3.4 Stochastic discount factor

The stochastic discount factor (or asset pricing kernel) is defined as

λj,t`k Pj,t
Mj,t|t`k :“ pβj qk Et
λj,t Pj,t`k

where λj,t :“ 1
Λj,t rhsdh and Λj,t rhs is the Lagrange multiplier associated with the individual
ş
nj Nj
household’s budget constraint. From here, one can establish the following relationships:

Mj,t|t “ 1
λj,t`1`k Pj,t`1
Mj,t`1|t`1`k “ pβj qk Et`1
λj,t`1 Pj,t`1`k
λj,t`1`k Pj,t λj,t`1 Pj,t λj,t`1`k Pj,t`1
Mj,t|t`1`k “ pβj qk`1 Et “ βj Et pβj qk Et`1
λj,t Pj,t`1`k λj,t Pj,t`1 loooooooooooooooomoooooooooooooooon
λj,t`1 Pj,t`1`k
“Mj,t`1|t`1`k

3.5 Utility maximization

Household h maximizes its lifetime utility subject to the budget constraint and the law of motion for
capital. The relevant part of the optimization problem is framed by the following Lagrangian:
#
8 C
ÿ k 1 ´ ¯1´σj
L “ Et pβj q µj,t`k C
Cj,t`k rhs ´ ϕC
j cj,t`k´1
k“0
1 ´ σj
+
χL
j ` s ˘1`σjL
´ L
Lj,t`k rhs
1 ` σj
#
8
ÿ Λj,t`k rhs
k
` Et pβj q Wj,t`k rhsLsj,t`k rhs ` Rj,t`k
K
Kj,t`k´1 rhs
k“0
Pj,t`k

´ Pj,t`k Cj,t`k rhs ´ Pj,t`k Ij,t`k rhs

` p1 ` ρB B B
j Pj,t`k qBj,t`k´1 rhs ´ Pj,t`k Bj,t`k rhs
ÿ B
` Bj,t`k´1 rs, hs ´ Pj,t`k rs1 sBj,t`k rs1 , hs
s1 PS
ÿ n
ÿ
` Eij,t`k pRij,t`k´1 Bij,t`k´1 rhs ´ Bij,t`k rhsq ´ Pj,t`k ΦB
ij,t`k rhs
i i
+
` other terms
# +
8
ÿ Λj,t`k rhsQK
k j,t`k rhs K
´
K
¯
` Et pβj q p1 ´ δj qKj,t`k´1 rhs ` 1 ´ Φj,t`k rhs Ij,t`k rhs ´ Kj,t`k rhs
k“0
Pj,t`k

where βjk Λj,t`k rhs and βjk Λj,t`k rhsQK

j,t`k rhs are the discounted Lagrange multipliers corresponding to
period pt ` kq’s real budget constraint and capital accumulation equation, respectively.

First-order condition with respect to domestic state-contingent bonds Bj,t rs, hs

Λj,t rhs 1 Pj,t

Et “ βj B Et
Λj,t`1 rhs
looooomooooon Pj,t rss Pj,t`1
λj,t
“Et λ
j,t`1

10
where s is the realized state in period t. As the right-hand side is independent of h, the Lagrange
multiplier is identical for all households and one can drop the index h. More precisely, defining the
per-capita variable λj,t :“ n1j Nj Λj,t rhsdh and aggregating over all households yields:
ş

ż ż
1 1 Pj,t 1
Λj,t rhsdh “ βj B Et Λj,t`1 rhsdh
n j Nj Pj,t rss Pj,t`1 nj Nj
looooooooomooooooooon loooooooomoooooooon
“:λj,t “:λj,t`1

Therefore Λj,t rhs “ λj,t . Re-arranging the equation yields the canonical asset pricing equation:

B λj,t`1 Pj,t
Pj,t rss “ βj Et “ Mj,t|t`1
λj,t Pj,t`1

The gross return on the bond in realized state s is defined by:

1
Rj,t :“ B rss
“ Mj,t|t`1
Pj,t

First-order condition with respect to domestic government bonds Bj,t rhs

B
Pj,t Λj,t`1 rhs Pj,t 1
Et “ βj Et “ Mj,t|t`1 “
1` ρB B
j Pj,t`1 Λj,t rhs Pj,t`1
loooomoooon Rj,t
λj,t`1
λj,t

Optimal Domestic Bond Decision

optimal domestic state-contingent bond decision in j

λj,t`1 Rj,t
1 “ βj Et
λj,t Πj,t`1

optimal government bond decision in j

1 ` ρB B
j Pj,t`1
Rj,t “ Et B
Pj,t

First-order condition with respect to international bonds Bij,t rhs

BΦB
ij,t rhs
BBij,t rhs
hkkkkkkkkkkkkkkkkkkikkkkkkkkkkkkkkkkkkj
˜ ¸
n n n
Eij,t B n Bij,t rhs
Eij,t Eij,t`1
Λj,t rhs ` Λj,t rhs ϕij Eij,t ´ bij “ βj Et Λj,t`1 rhs Rij,t
Pj,t Pj,t Pj,t Pj,t`1
˜ ˜ ¸ ¸
n
Pj,t 1 Λj,t`1 rhs Pj,t Eij,t`1
ô Bij,t rhs “ n βj Et n Rij,t ´ 1 ` bij
Eij,t ϕB ij Λj,t rhs Pj,t`1 Eij,t
loooomoooon
λj,t`1
λj,t

The right-hand side is independent of h, such that the optimal choice to hold international bonds is
identical for all households and one can drop the index h. More precisely, defining real per-capita bond

11
holdings as bij,t :“ 1
Bij,t rhs{Pi,t dh and aggregating over all households yields:
ş
nj Nj

˜ ¸
n P
Eij,t n
λj,t`1 Pj,t Eij,t`1
ż
i,t 1 Bij,t rhs 1
dh “ B βj Et n Rij,t ´ 1 ` bij
Pj,t nj Nj Pi,t
looomooon ϕ ij λj,t Pj,t`1 Eij,t
loooooooooomoooooooooon
r
Eij,t bij,t
n
λj,t`1 Pj,t Eij,t`1
ô 1 ` ϕB
` r ˘
ij Eij,t bij,t ´ bij “ Rij,t βj Et n
λj,t Pj,t`1 Eij,t

Note that due to arbitrage, it is imposed that the expected return on international bonds must equal
the nominal risk-free rate of the issuer country:

Rij,t “ Ri,t

Uncovered interest rate parity The first-order condition with respect to international bonds can be
re-arranged to yield uncovered interest rate parity (UIP):

¸¨ ´ ¯˛
1 ` ϕB r b
˜
λj,t`1 Pj,t 1 n
Eij,t ij E ij,t ij,t ´ bij
Mj,t|t`1 “ βj Et “ “ Et n
˝ ‚
λj,t Pj,t`1 Rj,t Eij,t`1 Ri,t
˜ ¸ ˆ
n
Eij,t`1
˙
Rj,t ` B
` r ˘˘
ô Et n “ 1 ` ϕ ij E ij,t bij,t ´ bij
Eij,t Ri,t

This equation suggests that the differential in interest rates between two countries is expected to reflect
the expected change in the nominal exchange rate (plus a risk premium in form of portfolio carrying
costs). Namely, the currency of the country with the higher interest rate is likely to depreciate, while
the currency of the country with the lower interest rate is expected to appreciate.

Uncovered Interest Rate Parity

if i-bonds are traded: optimal choice of i-currency bonds (per-capita) in j

˜ ¸
n
Eij,t`1
ˆ ˙
Rj,t
1 ` ϕB
` ` r ˘˘
Et n “ ij Eij,t bij,t ´ bij
Eij,t Ri,t

if i-bonds are not traded: no trade of i-currency bonds (per-capita) in j

bij,t “ 0

´ n
Eij,t
¯
where in the codes the change in the nominal exchange rate, n
Eij,t´1 , is defined as a distinct variable.

First-order condition with respect to Cj,t rhs

ˆ ˙ ´1
˘´σjC λj,t σC
Λj,t rhs “ µj,t Cj,t rhs ´ ϕC ` ϕC
` j
j cj,t´1 ô Cj,t rhs “ j cj,t´1
µj,t

12
Integrating over h yields:
ż ˆ ˙ ´1
1 λj,t σjC ˘´σjC
` ϕC C
`
Cj,t rhsdh “ j cj,t´1 ô λj,t “ µj,t cj,t ´ ϕj cj,t´1
n j Nj µj,t
looooooooomooooooooon
cj,t

As Λj,t rhs is independent of h, all households choose the same consumption level in equilibrium and
one can re-express the equation in per-capita terms using cj,t :“ n1j Nj Cj,t rhsdh. Again, this is a
ş

manifestation of the state-contingent bonds, acting as a hedge for individual income risk.

Marginal Consumption Utility

marginal consumption utility (per-capita) in j

˘´σjC
λj,t “ µj,t cj,t ´ ϕC
`
j cj,t´1

First-order condition with respect to Kj,t rhs

˜ ¸
QK
j,t rhs Λj,t`1 rhs QK
j,t`1 rhs
K
Rj,t`1
“ βj Et p1 ´ δjK q `
Pj,t Λj,t rhs
loooomoooon Pj,t`1 Pj,t`1
λj,t`1
λj,t

Defining the real per-capita shadow price of capital (marginal Tobin’s Q) as qj,t 1
K :“
ş
nj Nj QK
j,t rhs{Pj,t dh
one can re-write the first-order condition in terms of per-capita variables by aggregating over h.

Optimal Capital Decision

optimal capital decision (per-capita) in j

K
` K
p1 ´ δjK q ` rj,t`1
K
˘
λj,t qj,t “ βj Et λj,t`1 qj,t`1

where the real return on capital is symbolized by rj,t

K :“ RK {P .
j,t j,t

First-order condition with respect to investment Ij,t rhs

BΦK
j,t rhs
ΦK rhs
j,t BIj,t rhs
hkkkkkkkkkkkkikkkkkkkkkkkkj hkkkkkkkkkkkkkkkkkkikkkkkkkkkkkkkkkkkkj
˜ ¸
QK ϕK
ˆ ˙2 ˆ ˙
j,t rhs j Ij,t rhs K Ij,t rhs 1
1“ 1´ ´ 1 ´ ϕj ´1 Ij,t rhs
Pj,t 2 Ij,t´1 rhs Ij,t´1 rhs Ij,t´1 rhs
˙˜ ¸
Λj,t`1 rhs QK
ˆ
j,t`1 rhs K Ij,t`1 rhs ´1
´ βj Et ϕj ´1 Ij,t`1 rhs Ij,t`1 rhs
Λj,t rhs
loooomoooon Pj,t`1 Ij,t rhs pIj,t rhsq2
looooooooooooooooooooooooooomooooooooooooooooooooooooooon
λj,t`1
BΦK rhs
λj,t j,t`1
BIj,t rhs

Given the existence of state-contingent bonds, the investment ratios, Ij,t rhs{Ij,t´1 rhs, align across all
households in equilibrium. By defining per-capita investment as ij,t :“ n1j Nj Ij,t rhsdh, it follows that
ş

Ij,t rhs{Ij,t´1 rhs “ ij,t {ij,t´1 for all t. Defining the real per-capita shadow price of capital (marginal

13
Tobin’s Q) as qj,t 1
j,t rhs{Pj,t dh one can re-write the first-order condition in terms of per-
K :“
ş
nj Nj QK
capita variables by aggregating over h.

Optimal Investment Decision

optimal investment decision (per-capita) in j

˜ ˙¸
ϕK
ˆ ˙2 ˆ ˙ˆ
K j ij,t ij,t ij,t
λj,t “ λj,t qj,t 1´ ´1 ´ ϕK
j ´1
2 ij,t´1 ij,t´1 ij,t´1
ˆ ˙ˆ ˙2
K ij,t`1 ij,t`1
` βj Et λj,t`1 qj,t`1 ϕK
j ´1
ij,t ij,t

Capital accumulation in per-capita terms Re-visiting the law of motion for capital, one can re-write
the equation in per-capita terms, kj,t :“ n1j Nj Kj,t rhsdh, because the average investment ratio is
ş

identical across all households in equilibrium.

Capital Accumulation

capital accumulation (per-capita) in j

˜ ˙2 ¸
ϕK
ˆ
j ij,t
kj,t “ p1 ´ δjK qkj,t´1 ` 1´ ´1 ij,t
2 ij,t´1

3.6 Labor supply and wage setting

Labor union ℏ P Nj synthesizes a continuum of differentiated types of labor services, Lj,t rℏ, hs, to com-
pile the composite labor index Lj,t rℏs, which is offered to the intermediate goods producers. Households
are assumed to be the only suppliers of differentiated labor services within a framework of monopolistic
competition. This allows them some level of control over the salary for their specific labor product
h P Nj . Yet, as these distinct labor services are blended into the labor bundle, the aggregation un-
folds in an environment of perfect competition with the aggregating firm ℏ taking the wages set by
monopolistic households to be exogenously determined and beyond its control.

3.6.1 Labor aggregation

Aggregation technology The following constant elasticity of substitution (CES) technology is avail-
able to labor union ℏ for aggregating the differentiated types of labor:
W
˛ Wϵj
ϵW
¨
ˆ ˙ 1 ż j ´1 ϵ ´1
j
1 ϵW
j ϵW
Lj,t rℏs “ ˝ pLj,t rℏ, hsq j dh‚
nj Nj

where ϵW
j ą 1 is the intratemporal elasticity of substitution between the differentiated labor varieties.

14
Cost minimization The labor union ℏ takes the posted wage Wj,t rhs of household h as given while
determining the optimal quantity of hiring labor variety Lj,t rℏ, hs. The firm minimizes its nominal
expenditures in compliance with the technological constraint. The cost minimization problem is framed
by the following Lagrangian:

ϵWj
ϵW
» fi
1 j ´1 ϵW ´1
# ˆ ˙ +
ż ż j
1 ϵW
j ϵW
L“ Wj,t rhsLj,t rℏ, hsdh ` Wj,t Lj,t rℏs ´ – pLj,t rℏ, hsq j dhfl
Nj nj Nj

where Wj,t is the ideal (cost-minimizing) nominal wage rate of labor bundle Lj,t rℏs.

First-order condition with respect to Lj,t rℏ, hs

ˆ ˙´ϵW
1 Wj,t rhs j
Lj,t rℏ, hs “ Lj,t rℏs
nj Wj,t

Total demand for labor types Aggregating over ℏ yields total demand for labor type h:
ż ˆ ˙´ϵW ż
Wj,t rhs j 1
Ldj,t rhs “ Lj,t rℏ, hsdℏ “ Lj,t rℏsdℏ
Nj Wj,t n j Nj
looooooooomooooooooon
lj,t

where lj,t is the per-capita aggregate labor index.

Aggregated ideal wage rate The zero-profit condition

ż
Wj,t Lj,t rℏs “ Wj,t rhsLj,t rℏ, hsdh
Nj

along with the first-order condition (demand curve), implies that the ideal nominal wage index is
defined by:
˜ ¸ 1
1´ϵW
ż
1 1´ϵW j
Wj,t “ pWj,t rhsq j dh
nj Nj

3.6.2 Wage setting

Each household acts as wage setter for its differentiated labor service Lsj,t rhs in monopolistically com-
petitive labor markets, subject to the constraint of supplying enough labor to meet the demand of
the labor unions: Lsj,t rhs ě Ldj,t rhs. It is assumed that wages are determined by staggered nominal
contracts. That is, the probability of optimally resetting wages in a given period is 1 ´ θjW :
$
&WĂj,t rhs with probability 1 ´ θjW
Wj,t rhs “
j,t´1 rhs with probability θj
%W W

15
where W
Ăj,t rhs is the optimal reset wage of household h. A household granted the opportunity to re-
negotiate their wage contract in period t strategically sets this wage to maximize the expected present
value of their utility until the next opportunity for price adjustment arises. The probability of being
k
stuck with the current wage for k periods is pθjW q .

Lagrangian for wage re-optimization The relevant part of the Lagrangian representing the dis-
counted expected utility of a household undergoing wage re-optimization in period t is:
# ˜˜ ¸´ϵWj
¸1`σjL ˜ ¸´ϵWj
+
8
ÿ k χL
j Wj,t rhs
Ă λj,t`k Ă WĂj,t rhs
L “ Et pθjW βj q ´ lj,t`k ` Wj,t rhs lj,t`k ` . . .
k“0
1 ` σjL Wj,t`k Pj,t`k Wj,t`k
loooooooooooomoooooooooooon loooooooooooomoooooooooooon
Lsj,t`k rhs“Ld
j,t`k rhs Lsj,t`k rhs“Ld
j,t`k rhs
#
8
ÿ k χL
j ϵW L
j p1`σj q 1`σjL Ă ´ϵW L
j p1`σj q
“ Et pθjW βj q ´ pW j,t`k q plj,t`k q pW j,t rhsq
k“0
1 ` σjL
+
λj,t`k ϵW 1´ϵW
` Ăj,t rhsq j ` . . .
pWj,t`k q j plj,t`k qpW
Pj,t`k

First-order condition with respect to W

Ăj,t rhs

# +
W L 8
χL
Ăj,t rhsq´ϵj p1`σj q´1 Et k ϵW L 1`σjL
ÿ j
ϵW
j p1 ` σjL qpW pθjW βj q pWj,t`k q j p1`σj q plj,t`k q
k“0
1 ` σjL
# +
8
´ϵW ÿ k λj,t`k ϵW
“ pϵW
j ´ 1qpW
Ăj,t rhsq j
Et pθjW βj q pWj,t`k q j plj,t`k q
k“0
Pj,t`k

W p1`σ L q
Ăj,t rhsq1`ϵj
Multiply by pW j
and re-arrange:

k L ϵW L 1`σjL
˜ ¸ 1W
ϵW Et 8 W j p1`σj q pl
ř
j ´1
´ ¯1`ϵW L
j σj k“0 pθj βj q χj pWj,t`k q j,t`k q Sj,t
Wj,t rhs
Ă “ “: 2W
ϵW Et 8 W k λj,t`k ϵW Sj,t
ř
k“0 pθj βj q Pj,t`k pWj,t`k q
j pl
j j,t`k q

Note that the right-hand side is independent of h, so W

Ăj,t rhs is the same across all households who can
re-negotiate their wage and one can drop the index h from the optimal reset wage.

Recursive numerator
8
1W
ÿ k ϵj W p1`σ L q L
Sj,t “ Et pθjW βj q χL
j pWj,t`k q
j plj,t`k q1`σj
k“0
8
0 ϵW L 1`σjL
ÿ k ϵj W p1`σ L q L
“ pθjW βj q χL
j pWj,t q
j p1`σj q pl
j,t q ` Et pθjW βj q χL
j pWj,t`k q
j plj,t`k q1`σj
k“1
8
ϵj W p1`σ L q L ÿ k`1 L W L L
“ χL
j pWj,t q
j plj,t q1`σj ` Et pθjW βj q χj pWj,t`k`1 qϵj p1`σj q plj,t`k`1 q1`σj
k“0
ϵW L 1`σjL
“ χL
j pWj,t q
j p1`σj q pl
j,t q

16
 8
ÿ k ϵW L 1`σjL
` θjW βj Et Et`1 pθjW βj q χL j pWj,t`1`k q
j p1`σj q pl
j,t`1`k q
k“0
looooooooooooooooooooooooooooooooooomooooooooooooooooooooooooooooooooooon
W 1
“Sj,t`1

W p1`σ L q
Dividing both sides by pPj,t qϵj j , one can re-arrange the equation in terms of real variables and
inflation rates:
1W
s j,t
hkkkkkkkkkkikkkkkkkkkkj
˜ 1W
¸ ˜ ¸ϵW L
j p1`σj q
Sj,t Wj,t L
“ χL
j plj,t q1`σj
ϵW L
pPj,t q j p1`σj q
looPmo
j,ton
wj,t
˜ ¸ϵW L ˜
j p1`σj q 1W
¸
Pj,t`1 Sj,t`1
` θjW βj Et W L
looPmo
j,ton pPj,t`1 qϵj p1`σj q
loooooooooooomoooooooooooon
“Πj,t`1 W 1
sj,t`1

where s1j,tW is the re-scaled auxiliary recursive sum variable. The above equation then becomes:

W p1`σ L q L W p1`σ L q
s1j,tW “ χL
j pwj,t q
ϵj j plj,t q1`σj ` θjW βj Et pΠj,t`1 qϵj j s1j,t`1
W

Recursive denominator
8
2W
ÿ k λj,t`k W
Sj,t “ Et pθjW βj q pWj,t`k qϵj plj,t`k q
k“0
Pj,t`k
8
0 λj,t W ÿ k λj,t`k W
“ pθjW βj q pWj,t qϵj lj,t ` Et pθjW βj q pWj,t`k qϵj plj,t`k q
Pj,t k“1
Pj,t`k
8
λj,t W ÿ k`1 λj,t`k`1 W
“ pWj,t qϵj lj,t ` Et pθjW βj q pWj,t`k`1 qϵj plj,t`k`1 q
Pj,t k“0
Pj,t`k`1
8
λj,t W ÿ k λj,t`1`k W
“ pWj,t qϵj lj,t ` pθjW βj qEt Et`1 pθjW βj q pWj,t`1`k qϵj plj,t`1`k q
Pj,t k“0
Pj,t`1`k
loooooooooooooooooooooooooooooooomoooooooooooooooooooooooooooooooon
2
W
Sj,t`1

W ´1
Dividing both sides by pPj,t qϵj , one can re-arrange the equation in terms of real variables and
inflation rates:
˜ 2W
¸ ˜ ¸ϵW
j
˜ ¸ϵW
j ´1
˜ 2W
¸
Sj,t Pj,t Wj,t Pj,t`1 Sj,t`1
W “ λj,t lj,t ` θjW βj Et W
pPj,t qϵj ´1 Pj,t looPmo
j,ton looPmo
j,ton pPj,t`1 qϵj ´1
looooooomooooooon looooooooomooooooooon
2W
“wj,t Πj,t`1 2W
sj,t sj,t`1

where s2j,tW is the re-scaled auxiliary recursive sum variable. The re-scaled denominator equation then
becomes:

W W ´1
s2j,tW “ λj,t pwj,t qϵj lj,t ` θjW βj Et pΠj,t`1 qϵj s2j,t`1
W

17
Optimal reset wage Re-expressing the first-order condition in terms of the re-scaled variables s1j,tW
and s2j,tW yields:

˜ ¸1`ϵW L
j σj 1W W p1`σ L q
1`ϵW L W
Ăj,t ϵW
j Sj,t {pPj,t qϵj j ϵW
j s1j,tW
pw
rj,t q j σj “ “ “ 2W
Pj,t ϵW
j ´1
2W
Sj,t {pPj,t qϵj
W ´1
ϵW
j ´ 1 sj,t

W
where the real reset wage is defined using lowercase letter w Pj,t .
Ăj,t
rj,t “

Evolution of aggregate wages Re-examining the ideal nominal wage index through the lens of Calvo’s
staggered wage setting mechanism reveals:
˙ W ˙ W
Wj,t 1´ϵj Wj,t rhs 1´ϵj
ˆ ż ˆ
1´ϵW 1
pwj,t q j “ “ dh
Pj,t n j Nj Pj,t
˙ W ˙ W
Wj,t rhs 1´ϵj Wj,t rhs 1´ϵj
ż ˆ ż ˆ
1 1
“ dh ` dh
nj optimizers Pj,t nj non´optimizers Pj,t
˙1´ϵW ˙ W
Wj,t´1 rhs 1´ϵj
ˆ ż ˆ
W Pj,t´1 1
j
W 1´ϵW
“ p1 ´ θj qpw
rj,t q j ` θj dh
Pj,t n j Nj Pj,t´1
looooooooooooooooomooooooooooooooooon
1´ϵW
pwj,t´1 q j

ˆ ˙1´ϵW ˆ ˙1´ϵW
w
rj,t j W wj,t´1 j
ô 1 “ p1 ´ θjW q ` θjW pΠj,t qϵj ´1
wj,t wj,t

Evolution of wage dispersion Similar logic can be applied to derive the evolution of the wage dis-
persion term which is defined as:
˙ W
Wj,t rhs ´ϵj
ż ˆ
˚ 1
wj,t :“ dh
nj Nj Wj,t
˙ W ˙ W
Wj,t rhs ´ϵj Wj,t rhs ´ϵj
ż ˆ ż ˆ
1 1
“ dh ` dh
nj optimizers Wj,t nj non´optimizers Wj,t
˜ ¸´ϵWj ˙´ϵW ˙ W
Wj,t´1 rhs ´ϵj
ˆ ż ˆ
W W
Ăj,t {Pj,t
W Pj,t´1 Wj,t´1 {Pj,t´1
j 1
“ p1 ´ θj q ` θj dh
Wj,t {Pj,t Pj,t Wj,t {Pj,t n j Nj Wj,t´1
loooooooooooooooomoooooooooooooooon
˚
wj,t´1
ˆ ˙´ϵW ˆ ˙ϵW
w
rj,t j W wj,t j
“ p1 ´ θjW q ` θjW pΠj,t qϵj ˚
wj,t´1
wj,t wj,t´1

18
 Recursive Calvo Wage Setting

optimal reset wage in j

˜ ¸˜ ¸
1`ϵW L ϵW
j s1j,tW
pw
rj,t q j σj “
ϵW
j ´1 s2j,tW

recursive numerator (per-capita) for reset wage in in j

W p1`σ L q L W p1`σ L q
s1j,tW “ χL
j pwj,t q
ϵj j plj,t q1`σj ` θjW βj Et pΠj,t`1 qϵj j s1j,t`1
W

recursive denominator (per-capita) for reset wage in j

W W ´1
s2j,tW “ λj,t pwj,t qϵj lj,t ` θjW βj Et pΠj,t`1 qϵj s2j,t`1
W

evolution of aggregate wage in j

ˆ ˙1´ϵW ˆ ˙1´ϵW
w
rj,t j W wj,t´1 j
1 “ p1 ´ θjW q ` θjW pΠj,t qϵj ´1
wj,t wj,t

evolution of wage dispersion in j

ˆ ˙´ϵW ˆ ˙ϵW
w
rj,t j W wj,t j
˚
wj,t “ p1 ´ θjW q ` θjW pΠj,t qϵj ˚
wj,t´1
wj,t wj,t´1

4 Firms

4.1 Retail goods

Firm f within country j (f P Nj ) is a distributor, channeling a single retail good, denoted as Yj,t rf s,
for domestic absorption: consumption demand, investment demand ,the use of intermediate inputs
in the production of differentiated intermediate wholesale goods and government expenditures.2 This
retail good is composed of both locally produced and internationally sourced wholesale good bundles,
aggregated under perfect competition as producer prices are set beyond the influence of firm f .

Aggregation technology The distributor’s production technology integrates an index of locally pro-
duced wholesale goods, denoted as Yjj,t rf s, along with indices of wholesale goods produced abroad,
symbolized by Yij,t rf s, where the first index i ‰ j specifies the country of production, distinct from
the importing country j. The process is assumed to be of the constant elasticity of substitution (CES)
form:

2
The analysis is simplified by disregarding differences in the composition of trade flows. Specifically, the same trade
specification and parametrization across consumption, investment, intermediate input goods, and government spending
is adopted. This aligns with the “absorption-based perspective on trade” (Erceg, Guerrieri, and Gust 2008), which
posits that the primary factor influencing real imports is domestic absorption. Conversely, exports are determined by
the level of absorption in foreign markets.

19
 ¨ ˛ σj
˜˜ ˙2 ¸ ˜ ¸¸ σj ´1 σj ´1
1 σj ´1 1 ˆ σj
σj σj
ÿ σj φij Yij,t rf s{Yjj,t rf s Yij,t rf s
Yj,t rf s “ ˝γjj pYjj,t rf sq γij 1´ ´1
˚ ‹
`
2 Yij,t´1 rf s{Yjj,t´1 rf s ⋊ ‚
i‰j
τij,t

σj ‰ 1 is the elasticity of substitution in the terms of trade, which affects how easily products from
different countries can be substituted for each other in the production process of country j. The
parameter γij represents preference weights that reflect the degree of trade integration and that collec-
tively sum to 1, with γjj “ 1 ´ i‰j γij . These weights are parametrized as the product of home-bias
ř

parameter Ωij and country size ni :

γij “ Ωij ni

Here, 0 ă Ωij ď 1 influences the level of trade integration beyond mere economic size ni . A value of
Ωij “ 1 suggests that imports from country i would simply mirror its share in the global economy.
Adjusting the home bias parameter allows for modeling of various factors that affect trade flows,
including geographic proximity, cultural ties, and trade agreements. Therefore, values of Ωij close to
zero suggest minimal to no trade relations (Cwik, Müller, and Wolters 2011).

The term τij,t

⋊
ě 1 represents iceberg trade frictions. These costs capture the inherent inefficiencies and
losses encountered in international trade. Specifically, they account for a portion of goods that do not
reach the importing country—similar to how an iceberg loses mass as it moves through warmer waters.
In this analogy, goods melt away during distribution, resulting in the quantity shipped becoming
⋊
1{τij,t ď 1 units upon arrival. It is assumed that this reduction, or melting, occurs at the retail goods
stage, based on the observation that trading costs faced by bulk wholesale shippers are typically much
lower due to higher international price arbitrage at the wholesale level (Obstfeld and Rogoff 2000).

Lastly, the term in blue highlights the adjustment costs that arise when changing the proportion of
imported goods relative to those produced locally. This factor, influenced by parameter φij ě 0, reflects
the challenges that countries face in adjusting their trade balances over the short term; it effectively
lowers the trade elasticity temporarily as immediate adjustments are costly in the short-run (Erceg,
Guerrieri, and Gust 2006).

Cost minimization Each distributor chooses the optimal bundles of differentiated wholesale goods to
minimize its costs of distributing these into the retail good. Let Pjj,t represent the ideal producer price
index for the locally produced index Yjj,t , and Pij,t the corresponding producer-price index of imports
Yij,t , all expressed in the currency of j.

The cost minimization problem is framed by the following Lagrangian:

# +
8
ÿ Λj,t`k rf s ÿ
L “ Et pβj qk Pjj,t`k Yjj,t`k rf s ` Pij,t`k Yij,t`k rf s
k“0
Pj,t`k i‰j
#
8
ÿ
` Et pβj qk Λj,t`k rf s Yj,t`k rf s
k“0
˛ σjσ´1
j
˙˙ σjσ´1
¨ +
1 σj ´1 1 ˆˆ ´ ¯2 ˙ ˆ
ř σj φij Yij,t`k rf s{Yjj,t`k rf s Yij,t`k rf s j
σj σj
´ ˝pγjj q pYjj,t`k q ` i‰j γij 1´ 2 Yij,t`k´1 rf s{Yjj,t`k´1 rf s
´1 ⋊
‚
τij,t`k

20
where Λj,t rf s is the same Lagrange multiplier as in the household’s optimization problem, because
Yj,t rf s is the retail good used by the households for domestic absorption.

First-order condition with respect to Yjj,t rf s

ˆ ˙ ˆ ˙ # ˆ ˙ ´1
Pjj,t σj 1
σj
1
σj ´ 1
0 “ Λj,t rf s ´ Λj,t rf s pYj,t rf sq pγjj q σj pYjj,t rf sq σj
Pj,t σj ´ 1 σj
ˆ ˙ ˆˆ ¯2 ˙ ˆ ˙˙ ´1
σj
ÿ 1
σj ´ 1 φij
´
Yij,t rf s{Yjj,t rf s Yij,t rf s
` pγij q σj 1´ 2 Yij,t´1 rf s{Yjj,t´1 rf s
´1 ⋊
i‰j
σj τij,t

ˆ ˙ˆ ˙ˆ ˙˜ ¸+
Yij,t rf s{Yjj,t rf s Yij,t rf s p´1q Yij,t rf s
ˆp´φij q ´1
Yij,t´1 rf s{Yjj,t´1 rf s Yij,t´1 rf s{Yjj,t´1 rf s pYjj,t rf sq2 ⋊
τij,t
ˆ ˙ #
σj 1
´ Et βj Λj,t`1 rf s pYj,t`1 rf sq σj
σj ´ 1
ˆ ˙ ˆˆ ¯2 ˙ ˆ ˙˙ ´1
σ
ÿ 1
σj ´ 1 φij
´
Yij,t`1 rf s{Yjj,t`1 rf s Yij,t`1 rf s j
pγij q σj 1´ 2 Yij,t rf s{Yjj,t rf s
´1 ⋊
i‰j
σj τij,t`1

ˆ ˙ˆ ˙˜ ¸+
Yij,t`1 rf s{Yjj,t`1 rf s Yij,t`1 rf s{Yjj,t`1 rf s Yij,t`1 rf s
ˆp´φij q ´1
Yij,t rf s{Yjj,t rf s Yij,t rf s ⋊
τij,t`1

In the absence of import-adjustment costs, φij “ 0, the above first-order condition simplifies to:
ˆ ˙ ˆ ˙ ˆ ˙
Pjj,t σj 1
σj
1
σj
σj ´ 1 ´1
0 “ Λj,t rf s ´ Λj,t rf s pYj,t rf sq pγjj q pYjj,t rf sq σj
Pj,t σj ´ 1 σj

First-order condition with respect to Yij,t rf s

ˆ ˙ ˆ ˙ #
Pij,t σj 1
σj
0 “ Λj,t rf s ´ Λj,t rf s pYj,t rf sq
Pj,t σj ´ 1
ˆ ˙ ˆˆ ¯2 ˙ ˆ ˙˙ ´1
σ
˜
1
σj σj ´ 1 φij
´
Yij,t rf s{Yjj,t rf s Yij,t rf s j
pγij q 1´ 2 Yij,t´1 rf s{Yjj,t´1 rf s
´1 ⋊
σj τij,t
˜ ˆ ˙2 ¸˜ ¸
φij Yij,t rf s{Yjj,t rf s 1
1´ ´1
2 Yij,t´1 rf s{Yjj,t´1 rf s ⋊
τij,t
ˆ ˙ˆ ˙˜ ¸
Yij,t rf s{Yjj,t rf s 1{Yjj,t rf s Yij,t rf s
´φij ´1
Yij,t´1 rf s{Yjj,t´1 rf s Yij,t´1 rf s{Yjj,t´1 rf s ⋊
τij,t
¸+

ˆ ˙ #
σj 1
σj
´ Et βj Λj,t`1 rf s pYj,t`1 rf sq
σj ´ 1
ˆ ˙ ˆˆ ¯2 ˙ ˆ ˙˙ ´1
σ
1
σj ´ 1 φij
´
Yij,t`1 rf s{Yjj,t`1 rf s Yij,t`1 rf s j
pγij q σj 1´ 2 Yij,t rf s{Yjj,t rf s
´1 ⋊
σj τij,t`1

ˆ ˙ˆ ˙ˆ ˙˜ ¸+
Yij,t`1 rf s{Yjj,t`1 rf s Yij,t`1 rf s{Yjj,t`1 rf s p´1q Yij,t`1 rf s
p´φij q ´1
Yij,t rf s{Yjj,t rf s 1{Yjj,t rf s pYij,t rf sq2 ⋊
τij,t`1

In the absence of import-adjustment costs, φij “ 0, the above first-order condition simplifies to:

ˆ ˙ ˆ ˙ ˆ ˙ˆ ˙ ´1 ˜ ¸
Pij,t σj 1 1
σ j ´ 1 Y ij,t rf s
σj
1
0 “ Λj,t rf s ´ Λj,t rf s pYj,t rf sq σj pγij q σj ⋊
τij,t ⋊
Pj,t σj ´ 1 σj τij,t

Wholesale good demands in per-capita terms Per-capita wholesale demands of j for goods produced
in i can be defined by integrating over mass Nj and dividing by nj “ |Nj |: yij,t :“ n1j Nj Yij,t rf sdf .
ş

21
Similarly, the per-capita retail good is defined as yj,t :“ 1
Yj,t rf sdf . The ratio of Lagrange multi-
ş
nj Nj
pliers, Λj,t`1 rf s{Λj,t rf s, is constant across firms and equal to the ratio in per-capita terms, λj,t`1 {λj,t
as shown in the household’s optimization problem. Given that firms encounter identical prices, the
first-order conditions show that the ratios of demand, Yjj,t rf s{Yj,t rf s and Yij,t rf s{Yj,t rf s, remain con-
stant across firms. Therefore, the same ratios hold for the per-capita variables as well and one can
omit the firm-specific subscript f in the first-order conditions.3 Re-arranging the above equations
then yields the demand curves for local and imported wholesale goods in per-capita terms and relative
prices.

Optimal Demand Schedules For Wholesale Goods

local demand for wholesale goods (per-capita) produced in j

# ˜˜ ˆ ˙2 ¸ ˜ ¸¸ ´1
σj
1
σj
1
σj
´1
σj
ÿ 1
σj
φij yij,t {yjj,t yij,t
pjj,t “ pyj,t q pγjj q pyjj,t q ` pγij q 1´ ´1 ⋊
i‰j
2 yij,t´1 {yjj,t´1 τij,t
ˆ ˙˜ 2
¸˜ ¸+
yij,t {yjj,t pyij,t {yjj,t q 1
ˆ φij ´1 ⋊
yij,t´1 {yjj,t´1 yij,t´1 {yjj,t´1 τij,t
# ˜˜ ˆ ˙2 ¸ ˜ ¸¸ ´1σj
λj,t`1 1 ÿ 1 φ ij y {y
ij,t`1 jj,t`1 y ij,t`1
´ βj Et pyj,t`1 q σj pγij q σj 1´ ´1 ⋊
λj,t i‰j
2 yij,t {yjj,t τij,t`1
˜ ¸˜ ¸+
ˆ ˙ 2
yij,t`1 {yjj,t`1 pyij,t`1 q {yjj,t`1 1
ˆ φij ´1 ⋊
yij,t {yjj,t yij,t τij,t`1

demand of j for imports of wholesale goods (per-capita) produced in i

˜˜ ˙2 ¸ ˜ ¸¸ ´1
ˆ σj
1
σj
1
σj φij yij,t {yjj,t yij,t
pij,t “ pyj,t q pγij q 1´ ´1
2 yij,t´1 {yjj,t´1 ⋊
τij,t
˜˜ ˆ ˙2 ¸ ˜ ¸ ˆ ˙ˆ ˙˜ ¸¸
φij yij,t {yjj,t 1 yij,t {yjj,t yij,t {yjj,t 1
ˆ 1´ ´1 ´ φij ´1
2 yij,t´1 {yjj,t´1 ⋊
τij,t yij,t´1 {yjj,t´1 yij,t´1 {yjj,t´1 ⋊
τij,t
˜˜ ˙2 ¸ ˜ ¸¸ ´1
ˆ σj
λj,t`1 1
σj
1
σj φij yij,t`1 {yjj,t`1 yij,t`1
` β j Et pyj,t`1 q pγij q 1´ ´1
λj,t 2 yij,t {yjj,t ⋊
τij,t`1
˙˜ ¸
pyij,t`1 q2 {yjj,t`1
ˆ ˙ˆ
yij,t`1 {yjj,t`1 1
ˆ φij ´1
yij,t {yjj,t pyij,t q2 {yjj,t ⋊
τij,t`1

In the absence of import-adjustment costs, φij “ 0 these equations simplify to:

yjj,t “ γjj ppjj,t q´σj yj,t

´ ¯1´σj
yij,t “ γij ppij,t q´σj yj,t τij,t
⋊

Ideal producer price index The evolution of the ideal producer price index (PPI) is derived from the
zero-profit condition
ÿ
Pj,t Yj,t rf s “ Pjj,t Yjj,t rf s ` pPij,t q Yij,t rf s,
i‰j

3
To see this, the first-order conditions can be re-arranged such that prices, which are independent of f , are on the left-
hand side and the ratios of demands on the right-hand side. Without import adjustment costs, this is most obvious
and one can directly integrate over all f and use the per-capita definitions.

22
because all expenditures on Yjj,t rf s and Yij,t rf s are used for Yj,t rf s, so that Pj,t is the ideal price index.4
Integrating over all f and dividing by Pj,t yields:

1
ż
Pjj,t 1
ż ÿ ˆ Pij,t ˙ 1 ż
Yj,t rf sdf “ Yjj,t rf sdf ` Yij,t rf sdf
nj Nj looPmo
j,ton nj Nj Pj,t nj Nj
loooooooomoooooooon looooooooomooooooooon i‰j looomooon looooooooomooooooooon
yj,t pjj,t yjj,t pij,t yij,t

Ideal Producer Price Indices (Zero-Profit Condition)

zero-profit condition for distributed wholesale goods (per-capita) in j

ÿ
yj,t “ pjj,t yjj,t ` pij,t yij,t
i‰j

In the absence of import adjustment costs, one can also derive an intuitive representation by directly
inserting the demand schedules and re-arranging terms:
ÿ ´ ¯1´σj
1 “ γjj ppjj,t q1´σj ` γij ppij,t q1´σj τij,t
⋊

i‰j

The zero-profit condition is more general though.

4.2 Wholesale goods

Import and export price The import price Pji,t is denominated in the currency of the importing
country i. The connection between the import and export price Pji,t is determined by the currency
P CP,EX

pricing paradigm employed by the monopolists of producer country j:

Producer Currency Pricing

hkkkkkkkkkkkkkkikkkkkkkkkkkkkkj
n P CP,EX
ji,ton “ Eji,t P
loPomo ji,t
looomooon ,
i currency j currency

where Eji,t
n is the bilateral nominal exchange rate between countries j and i, converting the currency of

country j into the currency of country i. Note that we do not switch the order of indices in the export
price, but stick to the convention of exporting country first and importing country second.

Export margin It proves useful to conceptualize the export margin of country j versus i in terms
of relative prices. Specifically, this margin is defined by comparing the specific export price of goods
produced in j and sold to i against the price for local sales within country j itself:

Producer Currency Pricing

hkkkkkkkkkkkikkkkkkkkkkkj
P CP,EX
Pji,t
EM ji,t :“
Pjj,t

4
Alternatively, one can insert the optimal demand schedules into the aggregation technology. This is isomorphic to the
derivation via the zero-profit condition.

23
It can also be expressed in terms of the bilateral real exchange rate and relative producer prices:
pji,t
hkkkikkkj
ˆ ˙
n Pi,t Pji,t {Pi,t
EM ji,t :“ Eij,t
Pj,t P
looooomooooon jj,t {Pj,t
looomooon
r
Eij,t pjj,t

Export Margin

overall export margin for goods produced in j and sold to i relative to local sales in j

r pji,t
EM ji,t “ Eij,t
pjj,t

4.2.1 Export good aggregation (PCP)

A firm f within country i (f P Ni ) synthesizes a continuum of differentiated intermediate wholesale

goods, Yji,t rf, ms, to compile the index of wholesale goods Yji,t rfs, which is offered to the retail firms.
These goods are specifically produced for sales to country i under the producer currency pricing (PCP)
paradigm. The producers of the differentiated products are called export monopolists and they operate
within country j within a framework of monopolistic competition. This allows them some level of
control over the pricing of their specific variety m P Nj . Yet, as these distinct varieties are blended into
the wholesale goods bundle, the aggregation unfolds in an environment of perfect competition, with
the aggregating firm f taking the prices set by monopolists to be exogenously determined and beyond
its control.

Aggregation technology The following constant elasticity of substitution (CES) technology is avail-
able to firm f for aggregating the differentiated varieties:

ϵP
i
1 ϵP
˜ˆ ¸
˙ ż i ´1 ϵP ´1
1 ϵP
i
` ˘ ϵP
i
Yji,t rfs “ Yji,t rf, ms i dm
nj Nj

where ϵPi ą 1 is the intratemporal elasticity of substitution between the differentiated varieties.

Cost minimization The aggregator f, belonging to Ni , takes the listed price Pji,t rms of interme-
P CP,EX

diate goods from country j’s export branch as fixed input while determining the optimal quantity of
variety Yji,t rf, ms to purchase. The firm minimizes its nominal expenditures, which are converted to
its own currency i, in compliance with both the technological constraint and the producer currency
pricing (PCP) paradigm. The cost minimization problem is framed by the following Lagrangian:
$ ,
ϵP
i
’ «ˆ ˙1 ż P ff /
ϵi ´1 ϵP ´1 .
ż ’ /
n P CP,EX
& 1 ϵP
i
` ˘ ϵP
i
L“ Eji,t Pji,t rmsYji,t rf, ms dm ` Pji,t Yji,t rfs ´ Yji,t rf, ms i dm
Nj ’
’ nj Nj /
/
% -

24
where Pji,t is the ideal (cost-minimizing) price of one unit of the wholesale bundle Yji,t rfs, denominated
in i’s currency.

First-order condition with respect to Yji,t rf, ms

˜ P CP,EX
¸´ϵPi
n P
Eji,t rms
1 ji,t
Yji,t rf, ms “ Yji,t rfs
nj Pji,t

Total demand for intermediate PCP wholesale goods Aggregating over f yields total demand for
export variety m:
˜ P CP,EX
¸´ϵPi
n
Eji,t Pji,t
ż ż
d
rms ni 1
Yji,t rms :“ Yji,t rf, msdf “ Yji,t rfsdf
Ni Pji,t nj looooooooomooooooooon
ni Ni
yji,t

Aggregated ideal import wholesale price index The zero-profit condition

ż
n P CP,EX
Pji,t Yji,t rfs “ Eji,t Pji,t rms Yji,t rf, ms dm,
Nj

along with the first-order condition (demand curve), implies that the ideal wholesale producer price
index is defined by:
˜ ¸ 1
¯1´ϵPi 1´ϵP
ż
1 ´
n P CP,EX i
Pji,t “ Eji,t Pji,t rms dm
nj Nj

4.2.2 Local good aggregation

A firm f within country j (f P Nj ) synthesizes a continuum of differentiated intermediate wholesale

goods, Yjj,t rf, ms, to compile the index of wholesale goods Yjj,t rfs, which is offered to the retail firms.
The producers of the differentiated products are called local monopolists and they operate within
country j within a framework of monopolistic competition. This allows them some level of control
over the pricing of their specific variety m P Nj . Yet, as these distinct varieties are blended into the
wholesale goods bundle, the aggregation unfolds in an environment of perfect competition, with the
aggregating firm f taking the prices set by monopolists to be exogenously determined and beyond its
control.

Aggregation technology The following constant elasticity of substitution (CES) technology is avail-
able to firm f for aggregating the differentiated varieties:
P
˛ Pϵj
ϵP
¨
ˆ ˙1 ż j ´1 ϵ ´1
j
1 ϵPj ` ˘ ϵP
Yjj,t rfs “ ˝ Yjj,t rf, ms j dm‚
nj Nj

where ϵPj ą 1 is the intratemporal elasticity of substitution between the differentiated varieties.

25
Cost minimization The aggregator f, belonging to Nj , takes the listed price Pjj,t rms of interme-
diate goods from country j’s monopolists as fixed input while determining the optimal quantity of
variety Yjj,t rf, ms to purchase. The firm minimizes its nominal expenditures, in compliance with the
technological constraint. The cost minimization problem is framed by the following Lagrangian:
$ ,
ϵPj
’ /
ϵP
» fi
j ´1 ϵP ´1
’
’ ˆ ˙ 1
/
/
ż ż j
& 1 ϵP
j
` ˘ ϵP
.
L“ Pjj,t rmsYjj,t rf, ms dm ` Pjj,t Yjj,t rfs ´ – Yjj,t rf, ms j dmfl
Nj ’
’ nj Nj /
/
’
% /
-

where Pjj,t is the ideal (cost-minimizing) price of one unit of the wholesale bundle Yjj,t rfs, denominated
in the local currency.

First-order condition with respect to Yjj,t rf, ms

˜ ¸´ϵPj
1 Pjj,t rms
Yjj,t rf, ms “ Yjj,t rfs
nj Pjj,t

Total demand for intermediate local wholesale goods Aggregating over f yields total demand for
local variety m:
˜ ¸´ϵPj
Pjj,t rms
ż ż
d 1
Yjj,t rms :“ Yjj,t rf, msdf “ Yjj,t rfsdf
Nj Pjj,t nj Nj
looooooooomooooooooon
yjj,t

Aggregated ideal local wholesale price index The zero-profit condition

ż
Pjj,t Yjj,t rfs “ Pjj,t rms Yjj,t rf, ms dm,
Nj

along with the first-order condition (demand curve), implies that the ideal wholesale producer price
index is defined by:
˜ ¸ 1
1´ϵP
ż
1 ` ˘1´ϵPj j
Pjj,t “ Pjj,t rms dm
nj Nj

4.3 Intermediate goods

The intermediate goods sector consists of the following types of firms:

• m P Nj produces goods for the international market (export branch)

• m P Nj produces goods for the local market (local branch)

The international type adopts the concept of “pricing to market”, allowing producing firms the flexibility
to invoice their prices based on the export destinations. Operating under monopolistic competition,
these firms offer unique varieties and set prices on a staggered basis.

26
4.3.1 Production function

Each firm, regardless in which branch it operates, manufactures its distinct variety by blending a retail
bundle of intermediate inputs (Xj,t rms, Xj,t rms) with labor (Lj,t rms, Lj,t rms) and capital (Kj,t rms,
local EX local EX local

Kj,t rms) according to a Cobb-Douglas production function:

EX

´ ¯αX
j
´ αK 1´αK
¯1´αX
j
s,local local local j local j
Yj,t rms “ Aj,t Xj,t rms Kj,t rms Lj,t rms
´ ¯αX
j
´ αK 1´αK
¯1´αX
j
s,EX EX EX j EX j
Yj,t rms “ Aj,t Xj,t rms Kj,t rms Lj,t rms

αjX and αjK are the elasticities towards intermediate inputs and capital in production, respectively. It is
assumed that both labor as well as capital are internationally immobile. Aj,t is total factor productivity
which evolves according to an AR(1) process with a mean-reverting target Aj and persistence ρA
j .

Total Factor Productivity

total factor productivity in j:

logpAj,t q “ p1 ´ ρA A A
j q logpAj q ` ρj logpAj,t´1 q ` εj,t

4.3.2 Profit maximization

All firms belonging to the intermediate goods sector face the same gross nominal rental cost of capital
K , nominal wage index W
Rj,t j,t and price of retail good bundles Pj,t . They demand intermediate inputs,
capital and labor to maximize profits subject to the production function and to the constraint of
producing enough to meet local and foreign demands for their specific variety:

s,local s,EX
ÿ
d d
Yj,t rms ě Yjj,t rms Yj,t rms ě Yji,t rms
i‰j

It is assumed that firms are owned by households; accordingly, the stochastic discount factor is de-
rived from the optimal savings decision of the households and equal to Mj,t|t`k . Thus, the profit
maximization problem is framed by the following Lagrangian:
#
8
ÿ d local K local local
L “ Et Mj,t|t`k Pjj,t`k rmsYjj,t`k rms ´ Pj,t`k Xj,t`k rms ´ Wj,t`k Lj,t`k rms ´ Rj,t`k Kj,t`k rms
k“0
¯1´αX
˜ ¸+
´ ¯αX
j
´ αK 1´αK j
local local local j local j d
` M C j,t`k rms Aj,t`k Xj,t`k rms Kj,t`k rms Lj,t`k rms ´ Yjj,t`k rms
#
8 ÿ´ ¯
P CP,EX
ÿ d EX EX K EX
` Et Mj,t|t`k Pji,t`k rmsYji,t`k rms ´ Pj,t`k Xj,t`k rms ´ Wj,t`k Lj,t`k rms ´ Rj,t`k Kj,t`k rms
k“0 i‰j

¯1´αX
˜ ¸+
´ ¯ αX
j
´ αK 1´αK j ÿ´ d ¯
EX EX EX j EX j
` M C j,t`k rms Aj,t`k Xj,t rms Kj,t`k rms Lj,t`k rms ´ Yji,t`k rms
i‰j

where M C j,t rms and M C j,t rms denote the nominal marginal costs of the local and export branches,
local EX

respectively.

27
 local EX
First-order condition with respect to Xj,t rms, Xj,t rms

local ˆ ˙1´αX
Pj,t Xj,t rms ´
local
¯αX
j local αK
j local 1´αK
j
j
s,local
local
“ Aj,t Xj,t rms Kj,t rms Lj,t rms “ Yj,t rms
αjX M C j,t rms
EX ˆ ˙1´αX
Pj,t Xj,t rms ´
EX
¯αX
j EX αK
j EX 1´αK
j
j
s,EX
EX
“ Aj,t Xj,t rms Kj,t rms Lj,t rms “ Yj,t rms
αjX M C j,t rms

local EX
First-order conditions with respect to Kj,t rms, Kj,t rms

KK local ˆ ˙1´αX
Rj,t j,t rms
´
local
¯αX
j local αK
j local 1´αK
j
j
s,local
local
“ Aj,t Xj,t rms pKj,t rmsq pLj,t rmsq “ Yj,t rms
p1 ´ αjX qαjK M C j,t rms
KK EX ˆ ˙1´αX
Rj,t j,t rms
´
EX
¯αX
j EX αK EX 1´αK j
s,EX
EX
“ Aj,t Xj,t rms pKj,t rmsq j pLj,t rmsq j “ Yj,t rms
p1 ´ αjX qαjK M C j,t rms

local EX
First-order conditions with respect to Lj,t rms,Lj,t rms

local ¯αX ˆ ˙1´αX

Wj,t Lj,t rms ´
local j local αK
j local 1´αK
j
j
s,local
local
“ Aj,t Xj,t rms pKj,t rmsq pLj,t rmsq “ Yj,t rms
p1 ´ αjX qp1 ´ αjK qM C j,t rms
EX ¯αX ˆ ˙1´αX
Wj,t Lj,t rms ´
EX j EX αK
j EX 1´αK
j
j
s,EX
EX
“ Aj,t Xj,t rms pKj,t rmsq pLj,t rmsq “ Yj,t rms
p1 ´ αjX qp1 ´ αjK qM C j,t rms

4.3.3 Marginal costs

Expressing the first-order equations in terms of ratios:

local local local

Xj,t rms αjX Wj,t Xj,t rms αjX K
Rj,t Kj,t rms αjK Wj,t
“ X K
, “ , “
Llocal
j,t rms
p1 ´ αj qp1 ´ αj q Pj,t local
Kj,t rms p1 ´ αjX qαjK Pj,t Llocal
j,t rms
K K
p1 ´ αj q Rj,t
EX EX EX
Xj,t rms αjX Wj,t Xj,t rms αjX K
Rj,t Kj,t rms αjK Wj,t
“ , “ , “
EX
Lj,t rms X
p1 ´ αj qp1 ´ αjK q Pj,t EX
Kj,t rms X K
p1 ´ αj qαj Pj,t EX
Lj,t rms p1 ´ αjK q Rj,t
K

This shows that all firms, regardless of their type, set the same ratio of factor inputs as they face
the same input prices, production function and parameters. Accordingly, inserting into the first-order
conditions yields that marginal costs are the same for both the local and export branches, such that
one can drop the indices m and m. Defining M C j,t :“ 1{nj Nj M C j,t rms “ 1{nj Nj M C j,t rms yields:
ş local ş EX

ˆ ˙1´αX
1 K αj
K
p1´αK
j X
M C j,t “ ¯ X pRj,t q pWj,t q j q pPj,t qαj
1´αK 1´αj
´
αX αK
Aj,t pαjX q j p1 ´ αjX qpαjK q j p1 ´ αjK q j

4.3.4 Factor demands

Total factor demands are defined as the sum of local and export demands; hence, per-capita demands
for intermediate inputs xj,t , capital kj,t
d and labor ld can be used to combine the individual first-order
j,t

28
conditions in the following way:
˜ ż ¸
αjX
ż ż ż
1 local 1 EX Wj,t 1 local 1 EX
Xj,t rmsdm ` Xj,t rmsdm “ Lj,t rmsdm ` Lj,t rmsdm
nj Nj nj Nj p1 ´ αjX qp1 ´ αjK q Pj,t nj Nj nj Nj
looooooooooooooooooooooooooomooooooooooooooooooooooooooon looooooooooooooooooooooooooooomooooooooooooooooooooooooooooon
“:xj,t “:ld
j,t
˜ ¸
αjX K
ż ż ż ż
1 local 1 EX Rj,t 1 local 1 EX
Xj,t rmsdm ` Xj,t rmsdm “ K j,t rmsdm ` K j,t rmsdm
nj Nj nj Nj p1 ´ αjX qαjK Pj,t nj Nj nj Nj
looooooooooooooooooooooooooomooooooooooooooooooooooooooon loooooooooooooooooooooooooooooomoooooooooooooooooooooooooooooon
“:xj,t d
“:kj,t
˜ ¸
αjK
ż ż ż ż
1 local 1 EX Wj,t 1 local 1 EX
Kj,t rmsdm ` Kj,t rmsdm “ Lj,t rmsdm ` Lj,t rmsdm
nj Nj nj Nj p1 ´ αjK q Rj,t
K nj Nj nj Nj
looooooooooooooooooooooooooomooooooooooooooooooooooooooon looooooooooooooooooooooooooooomooooooooooooooooooooooooooooon
d
“:kj,t “:ld
j,t

That is, the ratios of factor demands are equally applicable to per-capita variables:

local EX local EX d local EX

xj,t Xj,t rms Xj,t rms xj,t Xj,t rms Xj,t rms kj,t Kj,t rms Kj,t rms
d
“ local “ EX , d
“ local “ EX , d
“ “
lj,t Lj,t rms Lj,t rms kj,t Kj,t rms Kj,t rms lj,t Llocal
j,t rms LEX
j,t rms

Consequently, the first-order conditions can be reformulated in per-capita terms.

Optimal Factor Demands

Optimal intermediate inputs demand (per-capita) in j:

˜ ¸αK X ˜
j p1´αj q
¸p1´αK X
j qp1´αj q
d
kj,t d
lj,t
1“ mcj,t αjX Aj,t
xdj,t xdj,t

optimal capital demand (per-capita) in j:

˜ ¸ αX ˜ ¸p1´αK X
j qp1´αj q
K
xdj,t j d
lj,t
rj,t “ mcj,t p1 ´ αjX qαjK Aj,t d d
kj,t kj,t

optimal labor demand (per-capita) in j:

˜ ¸αX ˜ ¸αK X
j p1´αj q
xdj,t j d
kj,t
wj,t “ mcj,t p1 ´ αjX qp1 ´ αjK qAj,t d d
lj,t lj,t

where rj,t
K :“ RK {P
j,t j,t denotes the real rental rate of capital, wj,t :“ Wj,t {Pj,t the real wage and
mcj,t :“ M C j,t {Pj,t are real marginal costs.

4.3.5 Export pricing (PCP)

Monopolists in the export market of country j, indexed by m P Nj , produce differentiated intermediate

goods designated for exporting to country i. The firms operate under the producer currency pricing
(PCP) paradigm, invoicing and setting the price Pji,t rms of its brand m in the domestic currency
P CP,EX

29
j. The demand for such intermediate goods is derived above and repeated here for convenience:
˜ P CP,EX
¸´ϵPi
n
Eji,t`k Pji,t`k rms
d ni
Yji,t`k rms “ yji,t`k
Pji,t`k nj

The export margin and the relationship between import and export prices under PCP are:
P CP,EX
Pji,t n P CP,EX
EM ji,t “ , Pji,t “ Eji,t Pji,t
Pjj,t

There is sluggish price adjustment due to the presence of nominal rigidities a la Calvo. That is, the
probability of optimally resetting prices in a given period is 1 ´ θjP :
$
P CP,EX
&PrP CP,EX rms
ji,t with probability 1 ´ θjP
Pji,t rms “
%Π P P CP,EX rms with probability θP
j ji,t´1 j

rms is the optimal reset price, decided in period t, of intermediate good m produced in j and
P CP,EX
Prji,t
sold to i in the currency of j (due to PCP). A firm granted the opportunity to adjust its price in
period t strategically sets this price to maximize the expected present value of its profits until the next
opportunity for price adjustment arises. If the firm cannot change its price for k periods, it only adjusts
for the inflation of country j, leading to an updated price of pΠj qk Pr rms in t ` k. The probability
P CP,EX
ji,t
k
of facing such a scenario is pθjP q . As firms are owned by the households, the stochastic discount factor
is derived from the household’s optimal consumption-savings decision and equal to Mj,t|t`k .

Lagrangian for price re-optimization The relevant part of the Lagrangian representing the discounted
sum of expected nominal profits for a firm undergoing price re-optimization in t is:
d
ji,t`k “Y rms
hkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkikkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkj
# ˜ P CP,EX
¸´ϵPi ˆ ˙+
8 ¯ En k
ÿ P k
´ P CP,EX
k ji,t`k P
rji,t rmspΠ jq ni
L “ Et pθj q Mj,t|t`k Prji,t rmspΠj q ´ M C j,t`k yji,t`k
k“0
Pji,t`k nj
˜ ¸´ϵPi ˆ # +
8 n
pΠj qk ´ P CP,EX ¯1´ϵPi ´ P CP,EX ¯´ϵPi
˙
ÿ P k Eji,t`k ni k
“ Et pθj q Mj,t|t`k yji,t`k pΠj q Pji,tr rms ´ M C j,t`k Pji,t r rms
k“0
Pji,t`k nj

P CP,EX
First-order condition with respect to Prji,t rms

˜ ¸´ϵPi ˆ
8 n
Eji,t`k pΠj qk
˙
P CP,EX ´ϵP ÿ k ni
pϵPi yji,t`k pΠj qk
i
´ 1qpPrji,t rmsq Et pθjP q Mj,t|t`k
k“0
Pji,t`k nj
˜ ¸´ϵPi
8 n
Eji,t`k pΠj qk
ˆ ˙
P CP,EX ´ϵP
i ´1
ÿ k ni
“ ϵPi pPrji,t rmsq Et pθjP q Mj,t|t`k M C j,t`k yji,t`k
k“0
Pji,t`k nj

ϵ P `1
Multiply by pPrji,t and re-arrange:
P CP,EX
rmsq i
´ n ¯´ϵPi ´ ¯
ř8 P k Eji,t`k Pji,t k ni 1
k“0 pθj q Mj,t|t`k pΠ q M C j,t`k nj yji,t`k
Et P CP,EX
ϵP
ˆ ˙
i ´1 P CP,EX
n P
Eji,t ji,t`k
j Sji,t
Prji,t rms “ “: 2
ϵP
i
ř8 P k
´ En
ji,t`k Pji,t k
¯´ϵPi ´
ni
¯
k
P CP,EX
Sji,t
Et k“0 pθi q Mj,t|t`k n P
Eji,t ji,t`k
pΠj q nj yji,t`k pΠj q

30
Note that the right-hand side is independent of m, so Prji,t rms for all m and one can drop
P CP,EX P CP,EX
“ Prji,t
the index m from the optimal reset price variable.
Recursive expression for the numerator
˜ n
¸´ϵP
8 Eji,t`k Pji,t i ˆ ˙
1P CP,EX ÿ k k ni
Sji,t “ Et pθjP q Mj,t|t`k n P
pΠj q M C j,t`k yji,t`k
k“0
Eji,t ji,t`k nj
˜ n P
¸´ϵP
Eji,t i ˆ ˙
0 ji,t ni
“ pθjP q Mj,t|t n P
pΠj q0 M C j,t yji,t
Eji,t ji,t nj
˜ n ¸´ϵP ˜ n
¸´ϵP
Eji,t`1 Pji,t i 8 Eji,t`k Pji,t`1 i ˆ ˙
ÿ k k ni
` n P
Et pθjP q Mj,t|t`k n
pΠj q M C j,t`k yji,t`k
Eji,t ji,t`1 k“1
Eji,t`1 Pji,t`k nj
ˆ ˙
ni
“ M C j,t yji,t
nj
˜ n ¸´ϵP ˜ n ¸´ϵP
Eji,t`1 Pji,t i 8 Eji,t`k`1 Pji,t`1 i ˆ ˙
ÿ
P k`1 k`1 ni
` n
Et pθj q Mj,t|t`k`1 n
pΠj q M C j,t`k`1 yji,t`k`1
Eji,t Pji,t`1 k“0
Eji,t`1 Pji,t`k`1 nj
˜ n
¸´ϵP
Eji,t`1 Pji,t i
ˆ ˙
ni λj,t`1 Pj,t P
“ M C j,t yji,t ` θjP βj Et n P
pΠj q´ϵi
nj λj,t Pj,t`1 Eji,t ji,t`1
˜ n ¸´ϵP
8 Eji,t`1`k Pji,t`1 i ˆ ˙
ÿ k k ni
ˆ Et`1 pθjP q Mj,t`1|t`1`k n
pΠ j q M C j,t`1`k yji,t`1`k
k“0
Eji,t`1 Pji,t`1`k nj
looooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooomooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooon
1P CP,EX
“Sji,t`1

˜ n
¸´ϵP
Eji,t`1 Pji,t i
ˆ ˙
1P CP,EX ni λj,t`1 Pj,t P 1P CP,EX
ô Sji,t “ M C j,t yji,t ` θjP βj Et n P
pΠj q´ϵi Sji,t`1
nj λj,t Pj,t`1 Eji,t ji,t`1

Multiplying both sides with pnj {ni qpλj,t {Pj,t q, one can re-arrange the equation in terms of real variables
and relative prices only:
˙ ˆ ˆ ˙
1P CP,EX nj λj,t M C j,t
Sji,t “ λj,t yji,t
ni Pj,t
looooooooooomooooooooooon Pj,t
loooomoooon
1P CP,EX mcj,t
“sji,t
˜ ¸ϵPi ˜ ¸ϵPi
E n P
ˆ ˙
P ji,t ji,t`1 ´ϵP 1P CP,EX nj λj,t`1
` θj βj Et n pΠj q i Sji,t`1
Pji,t Eji,t`1 ni Pj,t`1
looooooooooooomooooooooooooon
loooooooooooooomoooooooooooooon
1P CP,EX
P
P CP,EX ϵi “sji,t`1
“pΠji,t`1 q

1P CP,EX
where sji,t is the re-scaled auxiliary recursive sum variable. Note that Pji,t “ Eji,t and
n P P CP,EX
ji,t
the PCP export price inflation is defined as Πji,t {Pji,t´1 . The above equation then
P CP,EX P CP,EX P CP,EX
“ Pji,t
becomes:
˜ P CP,EX
¸ϵPi
1P CP,EX Πji,t`1 1P CP,EX
sji,t “ λj,t mcj,t yji,t ` θjP βj Et sji,t`1
Πj

Recursive expression for the denominator

˜ n
¸´ϵP ˆ
8 Eji,t`k Pji,t i ˙
2P CP,EX ÿ k k ni
Sji,t “ Et pθjP q Mj,t|t`k n P
pΠj q yji,t`k pΠj qk
k“0
Eji,t ji,t`k nj
˜ n P
¸´ϵP ˆ
Eji,t i ˙
0 ji,t ni
“ pθjP q Mj,t|t n P
pΠj q0 yji,t pΠj q0
Eji,t ji,t nj

31
 ˜ n
¸´ϵP ˜ n
¸´ϵP ˆ
Eji,t`1 Pji,t i 8 Eji,t`k Pji,t`1 i ˙
ÿ k ni
` n P
Et pθjP q Mj,t|t`k n
pΠj qk yji,t`k pΠj qk
Eji,t ji,t`1 k“1
Eji,t`1 Pji,t`k nj
ˆ ˙
ni
“ yji,t
nj
˜ n
¸´ϵP ˜ n
¸´ϵP ˆ
Eji,t`1 Pji,t i 8 Eji,t`k`1 Pji,t`1 i ˙
ÿ k`1 k`1 ni
` n P
Et pθjP q Mj,t|t`k`1 n
pΠj q yji,t`k`1 pΠj qk`1
Eji,t ji,t`1 k“0
Eji,t`1 Pji,t`k`1 nj
˜ n
¸´ϵP
Eji,t`1 Pji,t i
ˆ ˙
ni λj,t`1 Pj,t P
“ yji,t ` θjP βj Et n P
pΠj q1´ϵi
nj λj,t Pj,t`1 Eji,t ji,t`1
˜ n ¸´ϵP ˆ
8 Eji,t`1`k Pji,t`1 i ˙
ÿ
P k k ni
ˆ Et`1 pθj q Mj,t`1|t`1`k n
pΠ j q y ji,t`1`k pΠj qk
k“0
Eji,t`1 Pji,t`1`k nj
loooooooooooooooooooooooooooooooooooooooooooooooooooooooooooomoooooooooooooooooooooooooooooooooooooooooooooooooooooooooooon
2P CP,EX
“Sji,t`1

˙˜ n ¸´ϵP
Eji,t`1 Pji,t i
ˆ ˙ ˆ
2P CP,EX ni λj,t`1 Pj,t P 2P CP,EX
ô Sji,t “ yji,t ` θjP βj Et Πj n P
pΠj q´ϵi Sji,t`1
nj λj,t Pj,t`1 Eji,t ji,t`1

Multiplying both sides with pnj {ni qλj,t and re-arranging:

˙ ˆ
2P CP,EX nj
Sji,t λj,t “ λj,t yji,t
ni
looooooooooomooooooooooon
2P CP,EX
“sji,t

˙˜ ¸ϵPi ˜ ¸ϵP
ˆ n
Eji,t Pji,t`1 i
ˆ ˙
Πj ´ϵP 2P CP,EX nj λj,t`1
` θjP βj Et n pΠj q i Sji,t`1
Πj,t`1 Pji,t Eji,t`1 ni
looooooooooooomooooooooooooon
loooooooooooooomoooooooooooooon
2P CP,EX
P
P CP,EX ϵi “sji,t`1
“pΠji,t`1 q

2P CP,EX
where sji,t is the re-scaled auxiliary recursive sum variable. Note that Pji,t “ Eji,t and
n P P CP,EX
ji,t
the PCP export price inflation is defined as Πji,t {Pji,t´1 . The above equation then
P CP,EX P CP,EX P CP,EX
“ Pji,t
becomes:

ˆ ˙˜ P CP,EX
¸ϵPi
2P CP,EX Πj Πji,t`1 2P CP,EX
sji,t “ λj,t yji,t ` θjP βj Et sji,t`1
Πj,t`1 Πj

1P CP,EX 2P CP,EX
Optimal reset price The recursive equations for sji,t and sji,t can be combined to yield the
optimal reset price:

ˆ P ˙ S 1P CP,EX nj λj,t
¨ ´ ¯ ˛ ¨ 1 ˛
P CP,EX ˆ P ˙ P CP,EX
Pji,t
r ϵi ji,t n P ϵi s
“ ˝ i j,t
‚“ ˝ ji,t ‚
2P CP,EX nj 2P CP,EX
P ´1
´ ¯ P ´1
P j,t ϵi S λ ϵi s
looomooon ji,t ni j,t ji,t
P CP,EX
pji,t
“r

where the relative prices are defined using lowercase letters:

P CP,EX P CP,EX P CP,EX P CP,EX

prji,t “ Prji,t {Pj,t , pji,t “ Pji,t {Pj,t , pji,t “ Pji,t {Pi,t

Evolution of aggregate prices Re-examining the ideal price index through the lens of Calvo’s stag-
gered pricing approach reveals:

32
 P ji,t
hkkkkkkikkkkkkj ˜ ¸ 1
¯1´ϵPi 1´ϵP
ż
P CP,EX 1 ´
P CP,EX i
E n P “ E n P rms dm
ji,t ji,t nj Nj
ji,t ji,t
˜ P CP,EX
¸1´ϵPi P CP,EX
¸1´ϵPi ˜
¯1´ϵPi Pji,t Pji,t
ż
´
P CP,EX 1 rms
ô pji,t “ “ dm
Pj,t optimizers nj
Pj,t
˜ P CP,EX ¸1´ϵPi
Pji,t
ż
1 rms
` dm
nj non´optimizers Pj,t
ż ˜ P CP,EX
¸1´ϵPi
´ ¯ P
P CP,EX 1´ϵi
´
P CP,EX
¯1´ϵPi 1 P j,t´1 P ji,t´1 rms
ô pji,t “ p1 ´ θjP q prji,t ` θjP Πj dm
n j Nj Pj,t Pj,t´1
ˆ ˙1´ϵPi ż ˜ P CP,EX ¸1´ϵPi
´ ¯ P
P CP,EX 1´ϵi
´
P CP,EX
¯1´ϵPi P j,t´1 1 Pji,t´1 rms
ô pji,t “ p1 ´ θjP q prji,t ` θjP Πj dm
Pj,t n j Nj Pj,t´1
loooooooooooooooooomoooooooooooooooooon
P
P CP,EX 1´ϵi
“ppji,t´1 q
˜ P CP,EX
¸1´ϵPi ˜ P CP,EX
¸1´ϵPi
prji,t Pj,t´1 Pji,t´1 Pj,t
ô 1 “ p1 ´ θjP q
P CP,EX
` P CP,EX
θjP Πj
pji,t Pji,t Pj,t Pj,t´1
˜ P CP,EX ¸1´ϵPi ˜ ¸1´ϵPi
p ji,t Π j
ô 1 “ p1 ´ θjP q P CP,EX ` θjP
r
P CP,EX
pji,t Πji,t
˜ P CP,EX
¸1´ϵPi ˜ ¸1´ϵPi
P r
p
r ji,t P Πj
ô 1 “ p1 ´ θj q Eji,t ` θj P CP,EX
pji,t Π ji,t

Evolution of price dispersion Similar logic can be applied to derive the evolution of the price dis-
persion term which is defined as:
ż ˜ n P CP,EX ¸´ϵPi ż ˜ n P CP,EX ¸´ϵPi
1 E P
ji,t ji,t rms 1 E P rms
P CP,EX
p˚ji,t :“ dm “ ji,t ji,t dm
P CP,EX
n j Nj Pji,t n j Nj n P
Eji,t
 ji,t
˜ P CP,EX ¸´ϵPi ˜ P CP,EX ¸´ϵPi
Pji,t Pji,t
ż ż
1 rms 1 rms
“ P CP,EX
dm ` P CP,EX
dm
nj optimizers P nj non´optimizers P
ji,t ji,t

˜ P CP,EX
¸´ϵPi ˜ P CP,EX
¸´ϵPiż ˜ P CP,EX ¸´ϵPi
Prji,t {Pj,t Pji,t´1 1 Pji,t´1 rms
“ p1 ´ θjP q P CP,EX
` θjP P CP,EX
Πj P CP,EX
dm
Pji,t {Pj,t Pji,t n j Nj Pji,t´1
looooooooooomooooooooooon loooooooooooooooooomoooooooooooooooooon
ϵP P CP,EX
P CP,EX
pΠji,t {Πj q i “p˚
ji,t´1

˜ P CP,EX
¸´ϵPi
P CP,EX
¸ϵPi ˜
˚P CP,EX
prji,t Π ji,t P CP,EX
ô pji,t “ p1 ´ θjP q P CP,EX ` θjP p˚ji,t´1
pji,t Πj
˜ P
¸´ϵi ˜ P CP,EX ¸ϵPi
P CP,EX
pji,t Πji,t P CP,EX
“ p1 ´ θjP q Eji,t
r
` θjP
r
p˚ji,t´1
pji,t Πj

33
 Recursive Price Setting (PCP)

export price inflation of PCP goods produced in j and sold to i

P CP,EX EM ji,t
Πji,t “ Πjj,t
EM ji,t´1

optimal PCP reset price for PCP goods produced in j and sold to i
1P CP,EX
ϵPi sji,t
ˆ ˙
P CP,EX
prji,t “ 2P CP,EX
ϵPi ´ 1 sji,t

recursive numerator (per-capita) for PCP reset price of goods produced in j and sold to i

˜ P CP,EX
¸ϵPi
1P CP,EX Πji,t`1 1P CP,EX
sji,t “ λj,t mcj,t yji,t ` θjP βj Et sji,t`1
Πj

recursive denominator (per-capita) for PCP reset price of goods produced in j and sold to i

ˆ ˙˜ P CP,EX
¸ϵPi
2P CP,EX Πj Πji,t`1 2P CP,EX
sji,t “ λj,t yji,t ` θjP βj Et sji,t`1
Πj,t`1 Πj

evolution of export price inflation of PCP goods produced in j and sold to i

˜ P CP,EX
¸1´ϵPi ˜ ¸1´ϵPi
prji,t Πj
1 “ p1 ´ θjP q Eji,t
r
` θjP P CP,EX
pji,t Πji,t

price dispersion of PCP goods produced in j and sold to i

˜ P CP,EX
¸´ϵPi ˜ P CP,EX
¸ϵPi
˚P CP,EX
prji,t Πji,t P CP,EX
pji,t “ p1 ´ θjP q Eji,t
r
` θjP p˚ji,t´1
pji,t Πj

4.3.6 Local pricing

Monopolists in the local market of country j, indexed by m P Nj , produce differentiated intermediate

goods designated for domestic use. The firms invoice and set the price Pjj,t rms of its brand m in
its own currency j. The demand for such intermediate goods is derived above and repeated here for
convenience:
˜ ¸´ϵPj
d
Pjj,t`k rms
Yjj,t`k rms “ yjj,t`k
Pjj,t`k

There is sluggish price adjustment due to the presence of nominal rigidities a la Calvo. That is, the
probability of optimally resetting prices in a given period is 1 ´ θjP :
$
&Pr
jj,t rms with probability 1 ´ θjP
Pjj,t rms “
j jj,t´1 rms with probability θj
%Π P P

34
Prjj,t rms is the optimal reset price, decided in period t, of intermediate good m that is domestically
produced for local sales in j. A firm granted the opportunity to adjust its price in period t strategically
sets this price to maximize the expected present value of its profits until the next opportunity for price
adjustment arises. If the firm cannot change its price for k periods, it only adjusts for the inflation
of country j, leading to an updated price of pΠj qk Prjj,t rms in t ` k. The probability of facing such a
k
scenario is pθjP q . As firms are owned by the households, the stochastic discount factor is derived from
the household’s optimal consumption-savings decision and equal to Mj,t|t`k .

Lagrangian for price re-optimization The relevant part of the Lagrangian representing the discounted
sum of expected nominal profits for a firm undergoing price re-optimization in t is:
d
jj,t`k “Y rms
hkkkkkkkkkkkkkkkkkkkikkkkkkkkkkkkkkkkkkkj
¸ P
¯ Pr rmspΠ qk ´ϵj
# ˜ +
8 ´
P k j
ÿ k jj,t
L “ Et pθj q Mj,t|t`k Prjj,t rmspΠj q ´ M C j,t`k pyjj,t`k q
k“0
Pjj,t`k
˜ ¸´ϵPj # +
8
ÿ P k pΠj qk k
´ ¯1´ϵPj ´ ¯´ϵPj
“ Et pθj q Mj,t|t`k pyjj,t`k q pΠj q Pjj,t rms
r ´ M C j,t`k Pjj,t rms
r
k“0
Pjj,t`k

First-order condition with respect to Prjj,t rms

˜ ¸´ϵPj
pΠj qk
8
´ϵP ÿ k
pϵPj ´ 1qpPrjj,t rmsq j
Et pθjP q Mj,t|t`k pyjj,t`k q pΠj qk
k“0
Pjj,t`k
˜ ¸´ϵPj
pΠj qk
8
´ϵP
j ´1
ÿ k
“ ϵPj pPrjj,t rmsq Et pθjP q Mj,t|t`k M C j,t`k pyjj,t`k q
k“0
Pjj,t`k

ϵ `1 P
Multiply by pPrjj,t rmsq j and re-arrange:
´ ¯´ϵPj
ř8 k Pjj,t k
pθjP q Mj,t|t`k
˜ ¸
ϵP Et Pjj,t`k pΠj q M C j,t`k pyjj,t`k q 1
Sjj,t
j ´1 k“0
Prjj,t rms “ “: 2
ϵP
j
ř8 k
´ Pjj,t k
¯´ϵPj
k Sjj,t
Et k“0 pθiP q Mj,t|t`k Pjj,t`k pΠ j q pyjj,t`k q pΠj q

Note that the right-hand side is independent of m, so Prjj,t “ Prjj,t rms for all m and one can drop the
index m from the optimal reset price variable.
Recursive expression for the numerator
˜ ¸´ϵP
8 Pjj,t j
1
ÿ k k
pθjP q Mj,t|t`k
` ˘
Sjj,t “ Et pΠj q M C j,t`k yjj,t`k
k“0
Pjj,t`k
˜ ¸´ϵP
j
0 Pjj,t 0
“ pθjP q Mj,t|t pΠj q M C j,t pyjj,t q
Pjj,t
˜ ¸´ϵP ˜ ¸´ϵP
Pjj,t j 8 Pjj,t`1 j
ÿ k k
pθjP q Mj,t|t`k
` ˘
` Et pΠj q M C j,t`k yjj,t`k
Pjj,t`1 k“1
Pjj,t`k

“ M C j,t pyjj,t q
˜ ¸´ϵP ˜ ¸´ϵP
Pjj,t j 8 Pjj,t`1 j
ÿ k`1
pθjP q pΠj qk`1
` ˘
` Et Mj,t|t`k`1 M C j,t`k`1 yjj,t`k`1
Pjj,t`1 k“0
Pjj,t`k`1
˜ ¸´ϵP
j
λj,t`1 Pj,t Pjj,t P
“ M C j,t pyjj,t q ` θjP βj Et pΠj q´ϵj
λj,t Pj,t`1 Pjj,t`1

35
 ˜ ¸´ϵP
8 Pjj,t`1 j
ÿ k
pθjP q Mj,t`1|t`1`k pΠj qk
` ˘
ˆ Et`1 M C j,t`1`k yjj,t`1`k
k“0
Pjj,t`1`k
loooooooooooooooooooooooooooooooooooooooooooooooooooooooomoooooooooooooooooooooooooooooooooooooooooooooooooooooooon
1
“Sjj,t`1

˜ ¸´ϵP
j
1 λj,t`1 Pj,t Pjj,t P
1
ô Sjj,t “ M C j,t pyjj,t q ` θjP βj Et pΠj q´ϵj Sjj,t`1
λj,t Pj,t`1 Pjj,t`1

Multiplying both sides with pλj,t {Pj,t q, one can re-arrange the equation in terms of real variables and
relative prices only:
ˆ ˙
1 λj,t M C j,t
Sjj,t “ λj,t yjj,t
Pj,t
looomooon Pj,t
loooomoooon
“s1jj,t mcj,t
¸ϵPj ˜
P jj,t`1 P
1 λj,t`1
` θjP βj Et pΠj q´ϵj Sjj,t`1
Pjj,t Pj,t`1
loooooomoooooon
loooooomoooooon
ϵP “s1jj,t`1
pΠjj,t`1 q j

where s1jj,t is the re-scaled auxiliary recursive sum variable. Note that the local producer price inflation
is defined as Πjj,t “ Pjj,t {Pjj,t´1 . The above equation then becomes:

ˆ ˙ϵPj
Πjj,t`1
s1jj,t “ λj,t mcj,t yjj,t ` θjP βj Et s1jj,t`1
Πj
Recursive expression for the denominator
˜ ¸´ϵP
8 Pjj,t j
2
ÿ k k
pθjP q Mj,t|t`k yjj,t`k pΠj qk
` ˘
Sjj,t “ Et pΠj q
k“0
Pjj,t`k
˜ ¸´ϵP
j
0 Pjj,t 0
“ pθjP q Mj,t|t pΠj q pyjj,t q pΠj q0
Pjj,t
˜ ¸´ϵP ˜ ¸´ϵP
Pjj,t j 8 Pjj,t`1 j
ÿ k k
pθjP q Mj,t|t`k yjj,t`k pΠj qk
` ˘
` Et pΠj q
Pjj,t`1 k“1
Pjj,t`k

“ pyjj,t q
˜ ¸´ϵP ˜ ¸´ϵP
Pjj,t j 8 Pjj,t`1 j
ÿ k`1 k`1
pθjP q yjj,t`k`1 pΠj qk`1
` ˘
` Et Mj,t|t`k`1 pΠj q
Pjj,t`1 k“0
Pjj,t`k`1
˜ ¸´ϵP
j
λj,t`1 Pj,t Pjj,t P
“ pyjj,t q ` θjP βj Et pΠj q1´ϵj
λj,t Pj,t`1 Pjj,t`1
˜ ¸´ϵP
8 Pjj,t`1 j
ÿ k
pθjP q Mj,t`1|t`1`k pΠj qk yjj,t`1`k pΠj qk
` ˘
ˆ Et`1
k“0
P jj,t`1`k
loooooooooooooooooooooooooooooooooooooooooooooooooooomoooooooooooooooooooooooooooooooooooooooooooooooooooon
2
“Sjj,t`1

˙˜ ¸´ϵP
j
Pjj,t
ˆ
2 λj,t`1 Pj,t P
2
ô Sjj,t “ pyjj,t q ` θjP βj Et Πj pΠj q´ϵj Sjj,t`1
λj,t Pj,t`1 Pjj,t`1

Multiplying both sides with λj,t and re-arranging:

2
Sjj,t λj,t “ λj,t yjj,t
loomoon
“s2jj,t

36
 ˙˜ ¸ϵP
Pjj,t`1 j
ˆ
Πj P
2
` θjP βj Et pΠj q´ϵj Sjj,t`1 λj,t`1
Πj,t`1 Pjj,t loooooomoooooon
loooooomoooooon 2
“sjj,t`1
ϵP
pΠjj,t`1 q j

where s2jj,t is the re-scaled auxiliary recursive sum variable. Note that the local producer price inflation
is defined as Πjj,t “ Pjj,t {Pjj,t´1 . The above equation then becomes:

ˆ ˙ˆ ˙ϵPj
Πj Πjj,t`1
s2jj,t “ λj,t yjj,t ` θjP βj Et s2jj,t`1
Πj,t`1 Πj

Optimal reset price The recursive equations for s1jj,t and s2jj,t can be combined to yield the optimal
reset price:
˜ ¸¨ 1 λ
˛ ˜ ¸˜ ¸
s1jj,t
j,t
Prjj,t ϵPj Sjj,t Pj,t ϵPj
“ ˝
2
‚“
looPmo
j,ton ϵPj ´ 1 Sjj,t λj,t ϵPj ´ 1 s2jj,t
pjj,t
“r

where the relative reset price is defined using lowercase letter prjj,t “ Prjj,t {Pj,t .

Evolution of aggregate prices Re-examining the ideal price index through the lens of Calvo’s stag-
gered pricing approach reveals:
˙ P ˙ P
Pjj,t 1´ϵj Pjj,t rms 1´ϵj
ˆ ż ˆ
` ˘1´ϵPj 1
pjj,t “ “ dm
Pj,t nj optimizers Pj,t
˙ P
Pjj,t rms 1´ϵj
ż ˆ
1
` dm
nj non´optimizers Pj,t
ż ˆ ˙1´ϵPj
` ˘1´ϵPj P
` ˘1´ϵPj P 1 Pj,t´1 Pjj,t´1 rms
ô pjj,t “ p1 ´ θj q prjj,t ` θj Πj dm
n j Nj Pj,t Pj,t´1
˙1´ϵPj ˙ P
Pjj,t´1 rms 1´ϵj
ˆ ż ˆ
P Pj,t´1 1
` ˘1´ϵPj P
` ˘1´ϵPj
ô pjj,t “ p1 ´ θj q prjj,t ` θj Πj dm
Pj,t n j Nj Pj,t´1
loooooooooooooooooomoooooooooooooooooon
1´ϵP
“ppjj,t´1 q j

˜ ¸1´ϵPj ˜ ¸1´ϵPj
p jj,t P P P
j,t j,t´1 jj,t´1
ô 1 “ p1 ´ θjP q ` θjP
r
Πj
pjj,t Pjj,t Pj,t Pj,t´1
˜ ¸1´ϵPj ˜ ¸1´ϵPj
P
prjj,t P Πj
ô 1 “ p1 ´ θj q ` θj
pjj,t Πjj,t

Evolution of price dispersion Similar logic can be applied to derive the evolution of the price dis-
persion term which is defined as:
ż ˜ ¸´ϵPj
1 P jj,t rms
p˚jj,t :“ dm
n j Nj Pjj,t

37
 ˜ ¸´ϵPj ˜ ¸´ϵPj
Pjj,t rms Pjj,t rms
ż ż
1 1
“ dm ` dm
nj optimizers Pjj,t nj non´optimizers Pjj,t

˜ ¸´ϵPj ˜ ¸´ϵPj ż ˜ ¸´ϵPj

Prjj,t {Pj,t P jj,t´1 1 P jj,t´1 rms
“ p1 ´ θjP q ` θjP Πj dm
Pjj,t {Pj,t Pjj,t n j Nj Pjj,t´1
looooooooomooooooooon looooooooooooooooomooooooooooooooooon
ϵP
j “p˚
jj,t´1
pΠjj,t {Πj q
˜ ¸´ϵPj ˆ ˙ϵPj
prjj,t Πjj,t
ô p˚jj,t “ p1 ´ θjP q ` θjP p˚jj,t´1
pjj,t Πj

Recursive Price Setting (local)

optimal local reset price for local goods

˜ ¸
ϵPj s1jj,t
prjj,t “
ϵPj ´ 1 s2jj,t

recursive numerator (per-capita) for local reset price

ˆ ˙ϵPj
Πjj,t`1
s1jj,t “ λj,t mcj,t yjj,t ` θjP βj Et s1jj,t`1
Πj

recursive denominator (per-capita) for local reset price

ˆ ˙ˆ ˙ϵPj
Πj Πjj,t`1
s2jj,t “ λj,t yjj,t ` θjP βj Et s2jj,t`1
Πj,t`1 Πj

evolution of producer price inflation of local goods

˜ ¸1´ϵPj ˜ ¸1´ϵPj
prjj,t Πj
1 “ p1 ´ θjP q ` θjP
pjj,t Πjj,t

price dispersion of local goods

˜ ¸´ϵPj ˆ ˙ϵPj
prjj,t Πjj,t
p˚jj,t “ p1 ´ θjP q ` θjP p˚jj,t´1
pjj,t Πj

5 Government

5.1 Monetary policy

In each country, the monetary authority adheres to a Taylor-type rule for setting the policy interest
rate. The rule includes feedback in terms of deviations of the CPI from its target and the deviation of

38
actual (per-capita) GDP from its steady-state level:
˛ R
˙ψjR,gdp 1´ρj
¨
ˆ ˙ ˆ ˙ρR ˆ ˙ψR,Π ˆ
Rj,t Rj,t´1 j Π j,t
j gdpj,t R
“ ˝ ‚ eεj,t ,
Rj Rj Πj gdpj

Here, Rj,t denotes the policy interest rate in country j at time t which aligns with the nominal interest
rate on domestic bonds. The parameters ψjR,Π and ψjR,gdp reflect the policy’s sensitivity to inflation
and GDP deviations, respectively. ρR
j captures the extent of interest rate smoothing and εj,t represents
R

a monetary policy shock. The target value of inflation is denoted by Πj , while the steady-state values
of the policy interest rate and per-capita GDP are symbolized by Rj and gdpj .

Monetary Policy

interest rate rule in j:

˛ R
˙ψjR,gdp 1´ρj
¨
ˆ ˙ ˆ ˙ρR ˆ ˙ψR,Π ˆ
Rj,t Rj,t´1 gdp
˝ Πj,t
j j R
j,t
“ ‚ eεj,t ,
Rj Rj Πj gdp j

5.2 Fiscal policy

The fiscal authority issues domestic government bonds to finance interest payments on outstanding
debt and total government spending, symbolized by Gj,t ě 0. It adjusts lump-sum transfers T Rj,t rℏs
in order to clear the fiscal budget.

Fiscal budget The nominal fiscal budget is given by:

ż ż ż
B
P jj,t Gj,t ` Pj,t
looomooon T Rj,t rhsdh “ Pj,t
Bj,t rhsdh ´ p1 ` ρB B
j Pj,t q
Bj,t´1 rhsdh
Nj
looooooooooomooooooooooon Nj Nj
looooooooooooooooooooooooooooooomooooooooooooooooooooooooooooooon
spending
transfers domestic government bonds

Here, it is assumed that Gj,t is composed of domestic retail goods only and hence the price index Pjj,t
is used. Per-capita government spending, gj,t “ Gj,t {nj , is determined by a fiscal rule that follows an
AR(1) process:
ˆ ˙ ˆ ˙
gj,t gj,t´1
log “ ρG
j log ` εG
j,t
gj gj

Here, ρG
j captures the extent of fiscal smoothing and εj,t represents a fiscal spending shock. Transfers
G

include a response to the market value of the debt-to-GDP ratio (parametrized by ψjT R,B ě 0) and an
autoregressive term to allow for serial correlation (parametrized by ρTj R ):
˜ ¸
ˆ ˙ ˆ ˙ B
Pj,t´1 bj,t´1
trj,t trj,t´1
log “ ρTj R log ´ p1 ´ ρTj R qψjT R,B log
trj trj gdpj,t´1 BjGDP

where BjGDP denotes the target value of the debt-to-GDP ratio trj transfers in steady state.

39
 Fiscal Policy

government spending rule in j:

ˆ ˙ ˆ ˙
gj,t gj,t´1
log “ ρG
j log ` εG
j,t
gj gj

transfer rule in j:
˜ ¸
ˆ ˙ ˆ ˙ B
Pj,t´1 bj,t´1
trj,t trj,t´1
log “ ρTj R log ´ p1 ´ ρTj R qψjT R,B log
trj trj gdpj,t´1 BjGDP

6 Market clearing

6.1 Capital market

Total capital demand from the intermediate goods sector must equal total capital supplied by house-
holds:
˜ ż ż ¸ ż
1 local 1 EX ! 1
K rmsdm ` K rmsdm “ Kj,t´1 rhsdh
nj Nj j,t nj Nj j,t n j Nj
loooooooooooooooooooooooooooomoooooooooooooooooooooooooooon looooooooooomooooooooooon
d
“:kj,t “:kj,t´1

Note that the timing convention on Kj,t´1 rhs is end-of-period.

6.2 Labor market

The supply of differentiated labor input h must match the demand for this labor by the labor unions:
Lsj,t rhs “ Ldj,t rhs. Aggregating over h yields total labor supply (in per-capita terms):

˙ W
Wj,t rhs ´ϵj
ż ż ż ˆ
1 1 1
Lsj,t rhsdh “ Ldj,t rhsdh “ lj,t dh
nj Nj nj Nj n j Nj Wj,t
looooooooooooooomooooooooooooooon
˚
wj,t

where wj,t
˚ is the wage dispersion term due to nominal wage rigidities. Furthermore, this supply needs

to align with the labor demand from the monopolists (in per-capita terms):
˜ ż ż ¸
1 1
d
lj,t “ Ld,local
j,t rmsdm ` Ld,P
j,t
CP
rmsdm ˚
“ lj,t wj,t
nj Nj nj Nj

40
6.3 Aggregation

Aggregated international portfolio costs As all households choose the same portfolio, Bij,t rhs “
Bij,t , total international portfolio costs are
˜ ¸2
ϕB
ż
ij n Pi,t Bij,t
ΦB
ij,t rhsdh “ nj Eij,t ´bij
Nj 2 Pj,t looPmo
looomooon i,ton
r
Eij,t bij,t

Aggregated dividends Total dividends paid to households in j stem from the monopolists:
˙ P ˙ P
Pjj,t rms ´ϵj Pjj,t rms ´ϵj
ˆ ˆ
local
DIV j,t rms“ Pjj,t rms yjj,t ´ M C j,t yjj,t
Pjj,t Pjj,t
˙ P
Pjj,t rms ´ϵj
ˆ
ϵP 1´ϵP
“ pPjj,t q yjj,t pPjj,t rmsq
j j ´ M C j,t yjj,t
Pjj,t
¸ P
¯ E n P EX,P CP rms ´ϵi n
˜
´
EX n n P CP,EX ji,t ji,t i
DIV ji,t rms“ E ij,t Eji,t Pji,t rms ´ M C j,t yji,t
looomooon Pji,t nj
n
Ejj,t
˜ EX,P CP
¸´ϵPi
n P
Eji,t
n ϵP nj n EX,P CP 1´ϵP
i ni ji,t rms
“ Eij,t pPji,t q i yji,t pEji,t Pji,t rmsq ´ M C j,t yji,t
nj nj Pji,t

Aggregating over the monopolists yields total dividends (in per-capita) for the local and export
branches, respectively:

1´ϵP
j
pPjj,t q
hkkkkkkkkkkkkkkkkkikkkkkkkkkkkkkkkkkj
ż ˜ ż ¸
1 local ϵP 1 1´ϵ P
DIV j,t rmsdm“ pPjj,t q j yjj,t pPjj,t rmsq j dm
n j Nj n j Nj
˜ ż ˆ ˙ P ¸
1 Pjj,t rms ´ϵj
´M C j,t yjj,t dm
nj Nj Pjj,t
loooooooooooooooooomoooooooooooooooooon
p˚
jj,t

1´ϵP
i
pPji,t q
hkkkkkkkkkkkkkkkkkkkkkkikkkkkkkkkkkkkkkkkkkkkkj
ż ˜ ż ¸
1 EX n ϵP n j 1 n P CP,EX 1´ϵPi
DIV ji,t rmsdm“ Eij,t pPji,t q i yji,t pE P rmsq dm
n j Nj nj nj Nj ji,t ji,t
ż ˜ n P CP,EX ¸´ϵPi
¨ ˛
ni 1 Eji,t Pji,t rms
´M C j,t yji,t ˝ dm‚
nj n j Nj Pji,t
looooooooooooooooooooooomooooooooooooooooooooooon
P CP,EX
p˚
ji,t

Dividing by the CPI of j and combining the local dividends and export dividends from all countries
defines total real dividends in per-capita terms of country j:
local
1
ż
DIV j,t rms ÿ 1 ż DIV EX
ji,t rms
div j,t :“ dm ` dm
nj Nj Pj,t n
i‰j j Nj
Pj,t

41
 ˜ ¸ ˜ ¸
Pjj,t M Cj,t ˚ ÿ
n Pi,t Pji,t M Cj,t ˚P CP,EX ni
“ ´ p yjj,t ` Eij,t ´ pji,t yji,t
looPmo
j,ton Pj,t jj,t
loomoon i‰j
P j,t
looomooon P
lo
o i,t
moo n P j,t
loomoon n j
pjj,t mcj,t r
Eij,t pji,t mcj,t
´ ¯ ÿ´ ¯n
r P CP,EX i
“ pjj,t ´ mcj,t p˚jj,t yjj,t ` Eij,t pji,t ´ mcj,t p˚ji,t yji,t
i‰j
n j

Aggregated labor income Aggregated real labor income (in per-capita) is equal to:
W
˜ ¸
Wj,t rhsLdj,t rhs pWj,t qϵj
ż ż
1 1 1´ϵW Wj,t
dh “ lj,t pWj,t rhsq j dh “ lj,t “ wj,t lj,t
nj Nj Pj,t Pj,t n j Nj Pj,t
looooooooooooooooomooooooooooooooooon
1´ϵW
pWj,t q j

where Ldj,t rhs “ Lsj,t rhs is imposed due to labor market clearing.

Aggregated capital income Aggregated real capital income (in per-capita) is equal to:
ż KK
Rj,t
1 j,t´1 rhs K
dh “ rj,t kj,t´1
nj Nj Pj,t

Gross Domestic Product (GDP) Using an income accounting approach, GDP is the sum of capital
income, labor income and dividends:

K
gdpj,t :“ wj,t lj,t ` rj,t kj,t ` div j,t

where gdpj,t is GDP per-capita.

GDP Definition
GDP (per-capita) via income accounting in j
ÿ´ ¯n
K
` ˘ r P CP,EX i
gdpj,t “ wj,t lj,t ` rj,t kj,t ` pjj,t ´ mcj,t p˚jj,t yjj,t ` Eij,t pji,t ´ mcj,t p˚ji,t yji,t
i‰j
n j

6.4 Bond market

Domestic state-contingent bonds As households have the same stochastic discount factor, state-
contingent bonds are in zero net-supply in equilibrium:
ż
Bj,t rhsdh “ 0
Nj

42
Government bonds The fiscal budget constraint determines the amount of government bonds. It is
given in real per-capita terms as follows:
ż ż ˆ ˙ ż
Pjj,t Gj,t 1 B 1 Bj,t rhs B B Pj,t´1 1 Bj,t´1 rhs
` T Rj,t rhsdh “ Pj,t dh ´p1 ` ρj Pj,t q dh
P n
j,ton loomo
loomo j on n j Nj n j Nj P j,t Pj,t
loooomoooon n j Nj Pj,t´1
loooooooooomoooooooooon looooooooomooooooooon looooooooooomooooooooooon
pjj,t gj,t trj,t bj,t pΠj,t q´1 bj,t´1

International bonds Market clearing implies that the net supply of international bonds is zero:
ÿż ÿ 1
ż
Bji,t rhs
0“ Bji,t rhsdh “ ni dh
i Ni i
ni Ni Pj,t
looooooooomooooooooon
bji,t

Bond Market Clearing

fiscal budget constraint (per-capita) in j

B
1 ` ρB B
j Pj,t
pjj,t gj,t ` trj,t “ Pj,t bj,t ´ bj,t´1
Πj,t

market clearing for international j bonds (per-capita)

ÿ
0“ ni bji,t
i

6.5 Aggregate demand

The final retail good is used for domestic absorption, which defines the resource constraint for the
economy in per-capita terms as follows:
ż ż ż ˜ż ż ¸
1 1 1 1 d,local d,P CP
Yj,t rf sdf “ Cj,t rhsdh ` Ij,t rhsdh ` Xj,t rmsdm ` Xj,t rmsdm
n j Nj n j Nj n j Nj nj Nj Nj
loooooooomoooooooon looooooooomooooooooon loooooooomoooooooon looooooooooooooooooooooooooooomooooooooooooooooooooooooooooon
yj,t cj,t ij,t xj,t
Pjj,t Gj,t ÿ 1 ż
` ` ΦB
ij,t rhsdh
P j,t n j n j Nj
loomoon loomoon i‰j looooooooomooooooooon
pjj,t gj,t B
´ ¯2
ϕ
ij,t r b
2
Eij,t ij,t ´bij

Resource Constraint (Domestic Absorption)

aggregate demand index (per-capita) in j

ÿ ϕBij,t
´ ¯2
r
yj,t “ cj,t ` ij,t ` xj,t ` pjj,t gj,t ` Eij,t bij,t ´ bij
i‰j
2

43
6.6 Aggregate supply

Aggregate supply Aggregating the production function of the local and export monopolists over the
continuum of firms m P Nj and m P Nj yields (in per-capita terms):
ż ż
s 1 s,local 1 s,EX
yj,t :“ Yj,t rmsdm ` Yj,t rmsdm
nj Nj n j Nj
ż ´ ¯αX ˆ ˙1´αX
1 local j local αKj local 1´αK j
j
“ Aj,t Xj,t rms Kj,t rms Lj,t rms dm
n j Nj
ż ´ ¯αX ˆ ˙1´αX
1 EX j EX αK
j EX 1´αK j
j
` Aj,t Xj,t rms Kj,t rms Lj,t rms dm
n j Nj
ż ˜ local ¸αX ˜ local ¸αK p1´αX j q
1 Xj,t rms j Kj,t rms j local
“ Aj,t Lj,t rmsdm
nj Nj Llocal j,t rms
local
Lj,t rms
X
α ˜ EX ¸αK p1´αX j q
˜ EX
¸
Xj,t rms j Kj,t rms j
ż
1 EX
` Aj,t EX EX
Lj,t rmsdm
nj Nj Lj,t rms Lj,t rms
˜ ¸αX ˜ ¸ K X
αj p1´αj q ˜ ¸
j d
kj,t
ż ż
xj,t 1 local 1 EX
“ Aj,t d L rmsdm ` L rmsdm
lj,t d
lj,t nj Nj j,t nj Nj j,t
looooooooooooooooooooooooooomooooooooooooooooooooooooooon
d
“lj,t

Note that in the last step the individual input ratios are substituted by the per-capita ones (as they
are identical). Thus, the per-capita aggregate supply index is given by:

¯αX ˆ ˙1´αX
j
´ K K
s j d αj d 1´αj
yj,t “ Aj,t xdj,t pkj,t q plj,t q

At the same time, the market clearing condition for local and exported intermediate goods implies:

s,local s,EX
ÿ
d d
Yj,t rms “ Yjj,t rms, Yj,t rms “ Yji,t rms
i‰j

Aggregating over the continuum of firms yields (in per-capita terms):

ż ż
s 1 s,local 1 s,EX
yj,t :“ Yj,t rmsdm ` Yj,t rmsdm
n j Nj n j Nj
ż ż
1 d
ÿ 1 d
“ Y rmsdm ` Yji,t rmsdm
nj Nj jj,t i‰j
n j N j

¸ P
n P P CP,EX rms ´ϵi
˙ P ˜
Pjj,t rms ´ϵj Eji,t
ż ˆ ÿ 1 ż
1 ji,t ni
“ dm yjj,t ` dm yji,t
n j Nj Pjj,t n j N j
P ji,t
i‰j loooooooooooooooooooomoooooooooooooooooooon
n j
looooooooooooooomooooooooooooooon
p˚ P CP,EX
jj,t p˚
ji,t
ÿ P CP,EX ni
“ p˚jj,t yjj,t ` p˚ji,t yji,t
i‰j
nj

Equating both expressions for yj,t

s yields the aggregate supply expression.

44
 Aggregate Production

aggregate supply index (per-capita) in j

¯αX ˆ ˙1´αX
´ j d αj
K
d 1´αj
K j ÿ P CP,EX ni
Aj,t xdj,t pkj,t q plj,t q “ p˚jj,t yjj,t ` p˚ji,t yji,t
i‰j
nj

6.7 Aggregate budget constraint

To close the model, one needs to aggregate the household’s budget constraint over h:

Wj,t rhsLsj,t rhs K

Rj,t
ż ż
1 1
dh ` Kj,t´1 rhsdh
n j Nj Pj,t Pmo
lo
o j,ton nj Nj
loooooooooooooomoooooooooooooon looooooooooomooooooooooon
K
rj,t
wj,t lj,t kj,t´1
ż ż
Pj,t´1 1 Bj,t´1 rhs B Bj,t rhs 1
` p1 ` ρB B
j Pj,t q dh ´Pj,t ` T Rj,t rhs
looPmo
j,ton nj Nj Pj,t´1
looooooooooomooooooooooon
Pj,t
loomoon n j Nj
loooooooomoooooooon
pΠj,t q´1 bj,t´1 bj,t trj,t
ż ˜ ż ¸
1 1 local 1 ÿż EX
` DIV j,t rmsdm ` DIV ji,t rmsdm dh
nj Nj n j Nj nj i‰j Nj
loooooooooooooooooooooooooooooooomoooooooooooooooooooooooooooooooon
divj,t
ż ż 1
1 1 B 1 Bj,t rs , hs Bj,t´1 rs, hs
ÿ
“ Cj,t rhsdh ` Ij,t rhsdh ` Pj,t rs s ´
n j Nj n j Nj s1 PS
Pj,t Pj,t
looooooooomooooooooon loooooooomoooooooon loooooooooooooooooooooomoooooooooooooooooooooon
cj,t ij,t “0 (zero-net supply)
˜ ż ż ¸
ÿ 1 ż
n Pi,t 1 Bij,t rhs Pi,t´1 1 Bij,t´1 rhs
ÿ
` Eij,t dh ´ R ij,t´1 dh ` ΦB
ij,t rhsdh
i
Pj,t
looomooon nj Nj Pi,t loomoon
looPmo
i,ton nj Nj Pi,t´1 n j N j
i‰j looooooooomooooooooon
loooooooooomoooooooooon R
looooooooooomooooooooooon
i,t´1
r
Eij,t bij,t pΠi,t q´1 bij,t´1 ϕB
ij 2
r b
pEij,t
2 ij,t ´bij q

Using per-capita definitions and zero-net supply of state-contingent bonds yields:

K
1 ` ρB B
j Pj,t B
wj,t lj,t ` rj,t kj,t´1 ` bj,t´1 ´ Pj,t bj,t ` trj,t ` div j,t
Πj,t
˜ ¸
ÿ R i,t´1
ÿ ϕBij 2
r r
“ cj,t ` ij,t ` Eij,t bij,t ´ bij,t´1 ` pEij,t bij,t ´ bij q
i
Πi,t i‰j
2

Note that according to the aggregated real fiscal budget constraint the blue terms are equal to ´pjj,t gj,t .

45
 Aggregate Budget (Closing Condition)

Aggregate budget constraint in j:

´ ¯ ÿ´ ¯n
i
K
wj,t lj,t ` rj,t kj,t´1 ´ pjj,t gj,t ` pjj,t ´ mcj,t p˚jj,t yjj,t ` r
Eij,t pji,t ´ mcj,t p˚,EX
ji,t yji,t
i‰j
n j
˜ ¸
ÿ R i,t´1
ÿ ϕBij 2
r r
“ cj,t ` ij,t ` Eij,t bij,t ´ bij,t´1 ` pEij,t bij,t ´ bij q
i
Πi,t i‰j
2

According to Walras law one must omit a budget constraint; however, in the codes, a dummy variable
is introduced in the budget constraint of a designated closing country. This variable is intended to
always equal zero; therefore, serving as a convenient method to verify the market clearing conditions
within the model.

46
7 War shock scenario

The Home economy is impacted by a war shock, which is modeled as an autoregressive process. This
shock triggers a sequence of events influencing the economy’s dynamics: trade costs rise due to increased
iceberg costs with Home, capital in Home is destroyed, and Home’s productivity declines. Moreover,
all governments increase their spending to finance additional military and defense expenditures.

War shock The war shock, denoted by ωt , is modeled to follow an autoregressive process of order
two (AR(2)) with persistence parameters ρ1 and ρ2 and an exogenous innovation εωt , which triggers
the war shock by assuming a value of 1 at the onset of war and 0 otherwise.

War Shock
war shock process

ωt “ ρω1 ωt´1 ` ρω2 ωt´2 ` εωt

To ensure the process remains stationary, the roots of the war shock process, denoted by ρωI and ρωII ,
are constrained to lie within the unit circle:

ρω1 “ ρωI ` ρωII , ρω2 “ ´ρωI ¨ ρωII

Iceberg trade costs with Home The war shock increases iceberg trade costs of all countries in their
trade relationship with Home according to the war shock scaled by ∆⋊ ě 0.

Iceberg Trade Costs

iceberg trade costs in warring trade relationships (i ‰ H)

⋊ ⋊
τiH,t “ τHi,t “ 1 ` ∆ ⋊ ωt

no iceberg trade costs in non-warring trade relationships (j ‰ H):

⋊ ⋊
τij,t “ τji,t “1

Total factor productivity reduction in Home The war shock impacts total factor productivity in
Home, AH,t , by reducing it according to the war shock scaled by ∆A
H.

Total Factor Productivity Reduction

total factor productivity in H:

logpAH,t q “ p1 ´ ρA A A
H q logpAH q ` ρH logpAH,t´1 q ´ ∆H ωt

H.
Capital destruction in Home The war shock destroys capital in Home, KH,t , scaled by a factor ∆K
This modification changes the capital accumulation equation and, accordingly, the optimal capital and

47
investment decisions in Home.

Capital Destruction

capital accumulation (per-capita) in H:

˜ ˜ ˙2 ¸ ¸
ϕK
ˆ
K iH,t K
kH,t “ p1 ´ δH qkH,t´1 ` 1´ H ´1 iH,t e´∆H ωt
2 iH,t´1

optimal capital decision (per-capita) in H:

´ ¯
K K K ´∆K
λH,t qH,t “ βH Et λH,t`1 qH,t`1 p1 ´ δH qe H ωt`1 ` rH,t`1
K

optimal investment decision (per-capita) in H:

˜ ˙2 ˙¸
ϕK
ˆ ˆ ˙ˆ
´∆K iH,t i H,t i H,t
λH,t “ e H ωt K
λH,t qH,t 1´ H ´ 1 ´ ϕK ´1
H
2 iH,t´1 iH,t´1 iH,t´1
iH,t`1 2
ˆ ˙ˆ ˙
´∆K ωt`1 K K iH,t`1
` βH Et e H λH,t`1 qH,t`1 ϕH ´1
iH,t iH,t

Military spending in all countries All countries increase their government spending to finance ad-
ditional military expenditures; however, the magnitude of the increase differs between Home and the
other countries.

Military Spending

government spending in H
ˆ ˙ ˆ ˙
gH,t gH,t´1
log “ ρG
H log ` ∆G
H ωt
gH gH

government spending in j ‰ H
ˆ ˙ ˆ ˙
gj,t G gj,t´1
log “ ρ log ` ∆G ωt
gj gj

48
References
Cwik, Tobias, Gernot J. Müller, and Maik H. Wolters (Apr. 2011). “Does Trade Integration Alter
Monetary Policy Transmission?” In: Journal of Economic Dynamics and Control 35.4, pp. 545–564.
doi: 10.1016/[Link].2010.11.006.

Eichenbaum, M. S., B. K. Johannsen, and S. T. Rebelo (Feb. 2021). “Monetary Policy and the Pre-
dictability of Nominal Exchange Rates”. In: The Review of Economic Studies 88.1, pp. 192–228. doi:
10.1093/restud/rdaa024.

Erceg, Christopher J., Luca Guerrieri, and Christopher Gust (Mar. 2006). “SIGMA: A New Open
Economy Model for Policy Analysis”. In: International Journal of Central Banking 2.1.

Erceg, Christopher J., Luca Guerrieri, and Christopher Gust (Aug. 2008). “Trade Adjustment and the
Composition of Trade”. In: Journal of Economic Dynamics and Control 32.8, pp. 2622–2650. doi:
10.1016/[Link].2007.09.015.

Georgiadis, Georgios and Ben Schumann (Nov. 2021). “Dominant-Currency Pricing and the Global
Output Spillovers from US Dollar Appreciation”. In: Journal of International Economics 133, pp. 1–
16. doi: 10.1016/[Link].2021.103537.

Gomes, S., P. Jacquinot, and M. Pisani (Sept. 2012). “The EAGLE. A Model for Policy Analysis of
Macroeconomic Interdependence in the Euro Area”. In: Economic Modelling 29.5, pp. 1686–1714.
doi: 10.1016/[Link].2012.04.002.

Gopinath, Gita et al. (2020). “Dominant Currency Paradigm”. In: American Economic Review 110.3,
pp. 677–719. doi: 10.1257/aer.20171201.

Obstfeld, Maurice and Kenneth Rogoff (Jan. 2000). “The Six Major Puzzles in International Macroe-
conomics: Is There a Common Cause?” In: NBER Macroeconomics Annual 15, pp. 339–390. doi:
10.1086/654423.

Woodford, Michael (Aug. 2001). “Fiscal Requirements for Price Stability”. In: Journal of Money, Credit
and Banking 33.3, p. 669. doi: 10.2307/2673890.

49