openeconomy

...
Extensions on the Basic Model
• Open economy (based on work of Adolfson-Laséen-LindéVillani and

Christiano-Trabandt-Walentin (CTW))
• Search and matching in labor market (based on work of Gertler-Sala-Trigari

and Christiano-Ilut-Motto-Rostagno, CTW).
• Financial frictions (based on work of Bernanke-Gertler-Gilchrist and

Christiano-Motto-Rostagno, CTW)
1
Basic Model
• Model of Example #5
– Household preferences:
∞
Ã !
X Nt1+ϕ
t
E0 β { u (Ct) − exp (τ t) , u (Ct) ≡ log Ct
t=0
1 + ϕ
– Aggregate resources and household intertemporal optimization:
Rt
Yt = p∗t AtNt, uc,t = βEtuc,t+1
π̄ t+1
– Law of motion of price distortion:
⎛ Ã ! ε−1
ε
⎞−1
ε−1 ε
1 − θ (π̄ t ) θπ̄
pt = ⎝(1 − θ)
∗
+ ∗t⎠ . (1)
1−θ pt−1
2
Basic Model ...
– Equilibrium conditions associated with price setting:
1 + Etπ̄ ε−1
t+1 βθFt+1 = Ft (2)
∙ ε−1 ¸ 1−ε
1
1− θπ̄ t
Ft = Kt (3)
1−θ
= intermediate good firm marginal cost
z }| {
=Wt
Pt by household optimization
z }| ϕ{
ε exp (τ t) Nt 1 − ψ + ψRt
(1 − ν t) + Etβθπ̄ εt+1Kt+1
ε−1 uc,t At
= Kt (4)
3
Extension to Small Open Economy
• Outline
– the equilibrium conditions of the open economy model
∗ system jumps from 5-6 equations in basic model to 16 equations in 16
variables!
∗ additional variables:
rate of depreciation, exports, real foreign assets, terms of trade, real exchange rate, respectively
z }| {
f x
st, xt, at , pt , qt ,
price of domestic consumption (now, c is a composite of domestically produced goods and imports)
z}|{
pct ,
price of imports consumption price inflation
z}|{
m,c
z}|{
pt , π ct ,
reduced form object to (i) achieve technical objective, (ii) correct a fundamental failing of open economy models
z}|{
Φt ,
closed economy variables
z }| {∗
Rt, π̄ t, Nt, ct, Kt, Ft, pt .
4
Extension to Small Open Economy ...
– computing the steady state
– the ‘uncovered interest parity puzzle’, and the role of Φt in addressing the
puzzle.
– summary of the endogenous and exogenous variables of the model, as well

as the equations.
– several computational experiments to illustrate the properties of the model.
5
• Modifications to basic model to create open economy
– unchanged:
∗ household preferences
∗ production of (domestic) homogeneous good, Yt (= Atp∗t Nt)
∗ three Calvo price friction equations
– changes:
∗ household budget constraint includes opportunity to acquire foreign
assets/liabilities.
∗ intertemporal Euler equation changed as a reduced form accommodation
of evidence on uncovered interest parity.
∗ Yt = Ct no longer true.
∗ introduce exports, imports, current account.
∗ exchange rate,
domestic money
St = domestic currency price of one unit of foreign currency = foreign money
.
6
Homogeneous Final
Domestic Good Consumption
Foreign
Homogeneous
good
Intermediate
Good Producers
Exporters
Labor Market Foreign

Buyers
• Monetary policy: three approaches

– Taylor
µ ¶ rule µ ¶ µ c ¶ µ ¶
Rt Rt−1 π t+1 yt+1
log = ρR log +(1 − ρR) Et[rπ log +ry log ]+εR,t,
R R π̄ c y
(5)
where (could also add exchange rate, real exchange rate and other things):
π̄ c ~target consumer price inflation
εR,t ~iid, mean zero monetary policy shock
yt ~Yt/At
Rt ~‘risk free’ nominal rate of interest
– Svensson-style policy that solves Ramsey problem with the following
preferences:
X∞ ³ h i´2 µ µ ¶¶2
j c c c c c 4 yt
Et β { 100 π t π t−1π t−2π t−3 − (π̄ t ) + λy 100 log
j=0
y
2 ¡ ¢2
+λ∆R (400 [Rt − Rt−1]) + λs St − S̄ }
– straight Ramsey policy that maximizes domestic social welfare.
8
• Household budget constraint

StAft+1 + PtCt + Bt+1
h i
≤ BtRt−1 + St Φt−1Rt−1 Aft + WtNt + T ransf ers and prof itst
f
• Domestic bonds
Bt ~beginning of period t stock of loans
Rt ~rate of return on bonds
• Foreign assets
Aft ~beginning-of-period t net stock of foreign assets
(liabilities, if negative) held by domestic residents.
ΦtRtf ~rate of return on Aft
Φt ~premium on foreign asset returns, over foreign risk free rate, Rtf
9
• optimality of household foreign asset decision (verify this by solving

Lagrangian)
utility cost of one unit of foreign currency=St units of domestic currency, which is St /Ptc units of Ct
z }| {
uc,tSt
Ptc
conversion into utility units
z }| {
= βEt uc,t+1
quantity of domestic consumption goods that can be purchased from the payoff of one unit of foreign currency
z }| {
foreign currency payoff next period from one unit of foreign currency today
z }| {
St+1 Rtf Φt
× c
Pt+1
or
St St+1Rtf Φt
c = βEt c
Pt Ct Pt+1Ct+1
or
1 st+1Rtf Φt St Ct
= βEt c , st ≡ , ct = . (6)
ct π t+1ct+1 exp (∆at+1) St−1 At
10
• Optimality of household domestic bond decision:
1 Rt
= βEt c
PtcCt Pt+1Ct+1
– after scaling:
1 Rt
= βEt c . (7)
ct π t+1ct+1 exp (∆at+1)
11
• Final domestic consumption goods, Ct

– produced by representative, competitive firm using:
∙ ¸ η η−1
c
1 ¡ ¢ (η c −1) 1 (ηc −1) c
Ct = (1 − ωc) ηc Ctd ηc + ωcηc (Ctm) ηc
Ctd ~one-for-one transformation on domestic homogeneous output good, price Pt

Ctm ~imported good, with price Ptm,c
Ct ~final consumption good, with price, Ptc
η c ~elasticity of substitution between domestic and foreign goods.
– Profit maximization leads to:
Ctd = (1 − ωc) (pct)ηc Ct
µ c ¶ηc
pt
Ctm = ωc m,c Ct.
pt
h i 1−η1
m,c 1−η
pct = (1 − ω c) + ωc (pt ) c
c
(8)
c m,c
Pt m,c Pt
pct ≡ , p ≡
Pt t Pt
12
– Ctm is produced by competitive firm, which converts foreign homogeneous

output one-for-one into Ctm.
∗ Setting price equal to marginal cost:

³ ´
Pt = StPt 1 − ψ + ψ Rt , Ptf ~foreign currency price of foreign good.
m,c f f f f
or,
f
St Pt
f ³ ´ real exchange rate ≡ Pc ³ ´
c
P S
t t tP z}|{ t
pm,c
t = c 1 − ψ f
+ ψ f f
Rt = pc
t qt f
1 − ψ f + ψ f Rt .
Pt Pt
(9)
– Consumption good inflation:
" # 1−η1
1−η c
Ptc Ptpct (1 − ωc) + ωc (pm,c
t )
c
π ct ≡ c = = π̄ t ¡ m,c ¢1−ηc . (10)

Pt−1 Pt−1pct−1 (1 − ω c) + ωc pt−1
13
• Exports, Xt
– foreign demand for exports
Ã !−ηf
Ptx f x −η f f
Xt = Y t = (pt ) Yt (11)
Ptf
Ytf ~foreign output, Ptf ~price of foreign good, Ptx ˜ price of export
– Xt is produced one-for-one using the domestic homogeneous good by a
representative, competitive producer. Equating price, StPtx, to marginal
cost:
StPtx = Pt (ν xRt + 1 − ν x) ,
where ν x = 1 if all inputs must be financed in advance. Rewriting
qtpxtpct = ν xRt + 1 − ν x, (12)
where
real exchange rate f terms of trade
z}|{
z}|{ StPt x Ptx
qt ≡ c , pt = f
Pt Pt
qt π ft St
= st c , st = . (13)
qt−1 πt St−1
14
• Clearing in domestic homogeneous goods market:
output of domestic homogeneous good, Yt

= uses of domestic homogeneous goods
goods used in production of final consumption, Ct

z}|{ exports
z}|{ z}|{
government
= Ctd + Xt + Gt
= (1 − ωc) (pct)ηc Ct + Xt + Gt.

• Substituting out for Yt:
Atp∗t Nt = (1 − ωc) (pct)ηc Ct + Xt + Gt,

or,
p∗t Nt = (1 − ω c) (pct)ηc ct + xt + gt, (14)
Ct Xt Gt
ct ≡ , xt ≡ , gt ≡ .
At At At
15
• Current Account
– equality of international demand and supply for currency:
currency flowing out of the country
z }| {
acquisition of new net foreign assets, in domestic currency units
z }| {
StAft+1 + expenses on importst
currency flowing into the country

z }| {
receipts from existing stock of net foreign assets
z }| {
= receipts from exportst + StRt−1Φt−1Aft
f
,
– The pieces:
³ ´ µ pc ¶ηc
expenses on importst = StPtf 1 − ψ f + ψ f Rtf ω c t
Ct
pm,c
t
receipts from exportst = StPtxXt.

– Current account:
³ ´ µ pc ¶ηc
StAft+1 + StPtf 1 − ψ f + ψ f Rtf ωc m,c t f
Ct = StPtxXt + StRt−1 Φt−1Aft .
pt
16
– Divide current account by PtAt :
StAft+1 StPtf ³ ´ µ pc ¶ηc S P x S R f

Φ Af
t t t t−1 t−1 t
+ 1 − ψ f + ψ f Rtf ωc m,ct
ct = xt + ,
PtAt Pt pt Pt PtAt
or, using (9):
µ ¶ηc f f
pct st R Φt−1 a
aft + pm,c
t ωc ct = pctqtpxtxt + t−1 t−1
, (15)
pm,c
t π̄ t exp (∆at)
where aft is ‘scaled real, domestic value of foreign assets’:
f
St A
aft = t+1
PtAt
17
• ‘Risk’ adjustments
³ ´
f f
Φt = Φ at , Rt , Rt, φ̃t = (16)
³ ³ ´ ³ ¡ ¢´ ´
exp −φ̃a aft − ā − φ̃s Rtf − Rt − Rf − R + φ̃t
φ̃a > 0, small and not important for dynamics
φ̃s > 0, important
φ̃t ~mean zero, iid.
• Discussion of φ̃a.
– φ̃a > 0 implies (i) if aft > ā, then return on foreign assets low and aft ↓; (ii)
if aft < ā, then return on foreign assets high and aft ↑
– implication: φ̃a > 0 is a force that drives aft → ā in steady state,
independent of initial conditions.
– logic is same as reason why steady state stock of capital in neoclassical
growth model is unique, independent of initial conditions.
– in practice, φ̃a is tiny, so that its only effect is to pin down aft in steady state
and not affect dynamics (see Schmitt-Grohe and Uribe).
18
• Discussion of φ̃t
– Captures, informally, the possibility that there is a shock to the required

return on domestic assets. Perhaps this could be a crude stand-in for a
‘sub-prime mortgage crisis’, because it implies that people require a higher
return on domestic assets if they are to hold them.
• Discussion of φ̃s.
– φ̃s is an important reduced form feature, designed to correct an important

flaw in models of international finance. It represents a quick fix for the
problem, not a substitute for a longer-run solution.
– to better explain this, it is convenient to first solve for the model’s steady
state.
19
• Steady state
– household intertemporal efficiency conditions:

" #
1 st+1Rtf Φt sRf Φ
0 = Et −β c , steady state: 1 = β c (17)
ct π t+1ct+1 exp (∆at+1) π
∙ ¸
1 1 Rt R
0 = Et −β , steady state: 1 = β (18)
ct ct+1 π ct+1 exp (∆at+1) πc
– assumption about foreign households:
Rf
1 = β f (19)
π
Ptf
π ft ≡ f
(exogenous)
Pt−1
20
– Taylor rule:
π c = π̄ c (central bank’s inflation target). (20)
– from (10):
c Pt
π = π̄ ≡ . (21)
Pt−1
– using price friction equilibrium conditions:
1−θπ̄ ε
1−θ
p∗ = ³ ε−1
ε , (no distortion if π̄ = 1, )
´ ε−1 (22)
1−θ(π̄)
1−θ
1 ε−1
F = ε−1
, (don’t allow π̄ βθ < 1) (23)
1 − βθπ̄
ε ϕ+1 ∗
ε−1 (1 − ν) exp (τ ) N p (1 − ψ + ψR) ε
K = , (βθπ̄ < 1) (24)
1 − βθπ̄ ε
∙ ¸1
ε−1 1−ε
1 − θπ̄
K = F . (25)
1−θ
21
– other steady state conditions:

h i 1−η1
pc = (1 − ωc) + ωc (pm,c)1−ηc
c
(26)
m,c c
¡ f f f
¢
p =p q 1−ψ +ψ R (27)
qpxpc = ν xR + 1 − ν x (28)
p∗N = (1 − ωc) (pc)ηc c + x + g (29)
µ c ¶ηc
f c
¡ f f f
¢ p c x sRf Φaf
a + p q 1 − ψ + ψ R ωc m,c c = p qp x + (30)
p π̄
x = (px)−ηf y f (31)
– 15 equations: (17)-(31), 15 unknowns:
pc, pm,c, q, px, c, x, af , R, Φ, π̄, K, F, p∗, πc, s

– for convenience, set exogenous variables, g, ā,
g = ηg y, ā = η ay, where y = p∗N.
22
– algorithm for solving for the steady state:

∗ p∗, F, K can be computed from (21), (22), (23) and (25).
∗ solve (24) for N.
∗ solve
g = η g p∗N
af = ā = η ap∗N.
∗ (18), (19) imply
Rf R
= c (32)
πf π
∗ steady state depreciation, s, can be computed from the inflation
differential:
qt → q implies (see (13)) sπ f = πc.
∗ (17), (18) imply
sRf Φ = R, (33)
or after multiplication by π f and rearranging,
Rf R
f
Φ = c
, so (see (32)) Φ = 1 and at = ā (see (16))
π π
23
– rest of the algorithm solves a single non-linear equation in a single unknown.

– set
ϕ̃ = pcq.
– use (27), (28), (26):
pm,c = ϕ̃Rν,∗
x Rx
p = ,
ϕ̃
h i 1−η1
pc = (1 − ωc) + ω c (pm,c)1−ηc
c
ϕ̃
q = c.
p
– solve the resource constraint, (29), for c in terms of x :
c x sRf af
p qp x + π̄ − af
c= ¡ ¢ ³ ´ηc .
pc
p q 1 − ψf + ψ R
c f f
ωc pm,c
24
– use the latter to substitute out for c in the current account, (30):
µ ¶ηc f f
f c
¡ f f f
¢ pc pcqpxx + sRπ̄ a − af
a + p q 1 − ψ + ψ R ωc
pm,c ¡ f f
¢ ³ pc ´ηc
pcq 1 − ψ + ψ Rf ωc pm,c
f f
c x sR a
= p qp x + ,
π̄
which can be solved linearly for x.
– evaluate (31) and adjust ϕ̃ until it is satisfied. In practice, we set ϕ̃ = 1 and

used (31) to define y f .
25
• Uncovered interest rate parity puzzle and Φbt
– subtract equations (17)

" and (18): #
f
Rt − st+1Rt Φt
Et c = 0. (34)
ct+1π t+1 exp (∆at+1)
– totally differentiate the object in square brackets, and evaluate in steady

state
Rt − st+1Rtf Φt dRt 1 h f f f
i
d c = c
− c
sR dΦt + sdRt + R dst+1
ct+1π t+1 exp (∆at+1) cπ cπ
R − sRf c
− 2 d [ct+1 π t+1 exp (∆at+1)] ,
c
[cπ ]
so that, taking into account (33), (34) is, to a first approximation:
xt − x
R̂t = Etŝt+1 + R̂tf + Φ̂t, x̂t = log (xt) − log (x) = .
x
26
– Note:
¡ f¢
R̂t = log Rt − log (R) ' rt − log R, R̂tf = log Rtf − log R ' rtf − log Rf
Rt ≡ 1 + rt, Rtf ≡ 1 + rtf ,
so that:
rt − log (R) = log St+1 − log St − log s + rtf − log Rf + Φ̂t.
It follows from:
µ ¶
R
log (R) − log s − log Rf = log = 0,
sRf
that
rt = Et log St+1 − log St + rtf + Φ̂t (35)
³ ´ ³ ¡ f ¢´
f f
Φ̂t = log Φt = −φ̃a at − ā − φ̃s rt − rt − r − r + φ̃t
which is our log-linear expansion of (34).
27
– Uncovered Interest Parity (UIP)

∗ Under UIP, Φ̂t ≡ 0 and
rt > rtf → there must be an anticipated depreciation (instantaneous appreciation) of the

currency for people to be happy holding the existing stock of net foreign assets.
∗ Consider the standard ‘UIP regression’ (φ̃a ' 0, φ̃t = 0), involving risk
free rate differentials: ³ ´
f
log St+1 − log St = α + β rt − rt + ut.
∗ Substitute out for log St+1 − log St from (35) and make use of the
fact that a (rational expectations) forecast error is orthogonal to date t
information, to obtain:
³ ´
cov log St+1 − log St, rt − rtf
β̂ = ³ ´ = 1 − φ̃s,
f
var rt − rt
28
∗ In data,
· β̂ ' −.75, so UIP rejected (that’s the UIP puzzle)

· φ̃s = 1.75 → β̂ = −0.75.
∗ VAR impulse responses by Eichenbaum and Evans (QJE, 1992)
· data: rt ↑ after monetary policy shock → log St+j falls slowly for
j = 1, 2, 3, ... .
· UIP puzzle: rt ↑ and expected appreciation of the currency represents

a double-boost to the return on domestic assets. On the face of
it, it appears that there is an irresistible profit opportunity. Why
doesn’t the double-boost to domestic returns launch an avalanche
of pressure to buy the domestic currency? In standard models,
this pressure produces a greater instantaneous appreciation in the
exchange rate, until the familiar UIP overshooting result emerges -
the pressure to buy the currency leads to such a large appreciation,
29
that expectations of depreciation emerge. In this way, UIP leads to

the counterfactual prediction that a higher rt will be followed (after
an instantaneous appreciation) by a period of time during which the
currency depreciates.
· model’s resolution of the UIP puzzle: when rt ↑ the return required

for people to hold domestic bonds rises. This is why the double-boost
to domestic returns does not create an appetite to buy large amounts
of domestic assets. Possibly this is a reduced form way to capture the
notion that increases in rt make the domestic economy more risky.
(However, the precise mechanism by which the domestic required
return rises - earnings on foreign assets go up - may be difficult
to interpret. An alternative specification was explored, with risk-
premia affecting domestic bonds, but this resulted in indeterminacy
problems.)
30
• Model dynamics
– 16 equations: price setting, (1), (2),(3) and (4); monetary policy rule, (5);
household intertemporal Euler equations (6), (7); relative price equations
(13), (8), (9), (10), (12); aggregate resource condition, (14); current account,
(15); risk term, (16); demand for exports (11).
– 16 endogenous variables: pct, pm,c

t , qt , Rt , π̄ t , π c x
,
t tp , Nt , p∗ f
t , at , Φ̂t , st, xt , ct , Kt , Ft .
– exogenous variables: Rtf , ytf , φ̃t, gt, εR,t, ∆at, τ t, π ft .

for the purpose of numerical calculations, these were modeled as
independent scalar AR(1) processes.
– the model was solved in the manner described above:

∗ compute the steady state using the formulas described above
∗ log-linearize the 16 equations about steady state
∗ solve the log-linearized system
∗ these calculations were made easy by implementing them in Dynare.
31
• Numerical examples
• Parameter values:
foreign and domestic inflation same no financial frictions small weight on aft in Φt
z }| { z }| { z }| {
π̄ c = π f = 1.005, ψ = ψ f = ν x = 0, φ̃a = 0.03 ,
prices unchanged on average for 1 year 1/ϕ Frisch elasticity
−1/4
z }| { z }| {
β = 1.03 , θ = 3/4 , ϕ=1 ,
subsidy extinguishes monopoly power in labor margin
modest elasticity of demand for domestic intermediate goods z }| {
z }| { ε−1
ε=6 , 1 − νt = ,
ε
elasticity of substitution between domestic and foreign inputs in producing final consumption
z }| {
ηc = 5
roughly 60% of domestic final consumption is composed of domestic content share of g in y net foreign assets/y
z }| { z }| { z }| {
ωc = 0.4 , η g = 0.3 , ηa = 0 ,
elasticity of demand for exports as function of relative price paid by foreigners monetary policy rule parameters
z }| { z }| {
η f = 1.5 , ρR = 0.9, rπ = 1.5, ry = 0.15
32
• iid shock, 0.01, to εR,t.

– φ̃s = 0 →after instantaneous appreciation, positive εR,t shock followed by
depreciation.
– for higher φ̃s, shock followed by appreciation.
– long run appreciation is increasing function of persistence of ρR.
Response of log exchange rate to monetary policy shock
0
φ =0
s
φs = 0.5
−0.01
φs = 1.0
φs = 1.75
−0.02
φs = 1.75, ρR = 0.8
−0.03
−0.04
−0.05
−0.06
−0.07
−0.08
−0.09
−0.1
0 1 2 3 4 5 6 7
33
We now consider a monetary policy shock, εR,t = 0.01. According to (5),

implies a four percentage point (at an annual rate) policy-induced jump in Rt.
The dynamic effects are displayed in the following figure, for φ̃s = 0, φ̃s = 1.75.
response to monetary policy shock
log, homogeneous output log, consumption domestic risk free rate (APR) log level, exchange rate
0
−0.07
−0.02 7
−0.05 −0.08
−0.04 6
−0.09
−0.1 −0.06 −0.1
5
0 10 0 10 0 10 0 10
log, hours worked consumer price inflation (APR) log, real exchange rate real foreign assets
0 0 0
−0.02
−0.01
−0.05 −10 −0.04
−0.02
−0.06
−0.03
−0.1 −20 −0.08
0 10 0 10 0 10 0 10
log, terms of trade log, exports log, imports

0.06 0 0.06
−0.02 0.04 φ =0
0.04 s
−0.04 0.02
0.02 −0.06 φs = 1.75
0
−0.08
0
0 10 0 10 0 10
Note: (i) appreciation smaller, though a more drawn out, when φ̃s is big; (ii)
smaller appreciation results in smaller drop in net exports, so less of a drop in
demand, so less fall in output and inflation; (iii) smaller drop in net exports results
in smaller drop in real foreign assets.
34
Consider now a domestic economy risk premium shock, a jump in the

innovation to φ̃t equal to 0.01.
response to country risk premium shock
log, homogeneous output x 10

−3 log, consumption
domestic risk free rate (APR) log level, exchange rate
5 5.8 0.08
0.06
5.6 0.07
0.04 0
5.4 0.06
0.02 −5
5.2
0 0.05
0 10 0 10 0 10 0 10
log, hours worked consumer price inflation (APR)log, real exchange rate real foreign assets
0.04 0.2
0.06 0.03
0.04 0.02 0.15
0.02
0.02 0.01
0.1
0 0
0
0 10 0 10 0 10 0 10
log, terms of trade log, exports log, imports

0 0.08 0
−0.02 0.06
−0.05
0.04
−0.04 0.02 −0.1
−0.06 0
0 10 0 10 0 10
With the reduced interest in domestic assets, (i) the currency depreciates, (ii)
net exports rise, (iii) hours and output rise, (iv) the upward pressure on costs
associated with higher output leads to a rise in prices.
35
Next we consider a 0.01 innovation in log, government consumption, gt.

response to government consumption shock
log, homogeneous output log, consumption log level, exchange rate

x 10
−3
x 10
−3 domestic risk free rate (APR) x 10
−3
1 4.95 −4.5
−1
4.94 −5
0
−1.5
4.93 −5.5
−1
−2 4.92
−6
0 10 0 10 0 10 0 10
−3
log, hours worked consumer
−3
price inflation (APR) log, real exchange rate
−4 −3
real foreign assets
x 10 x 10 x 10 x 10
1 0 −3
−0.5
−5 −4
0 −1
−5
−1.5 −10
−1 −6
−2 −15 −7
0 10 0 10 0 10 0 10
log, imports
log, terms of trade log, exports
−3 −3 −3
x 10 x 10 x 10
3 0
3
2 2
−2
1
1
0
−4
0
0 10 0 10 0 10
After a delay, the higher gt leads to a rise in output. However, there is so much
crowding out in the short run that output actually falls. There is crowding out of
net exports and consumption because of the effects created by a higher interest
rate. The higher interest rate directly reduces consumption, and by making
the currency appreciate, it produces a fall in net exports. The initial drop in
government spending in the wake of a rise in government spending is interesting.
36
Introducing Labor Market Search and Matching
Frictions into Open Economy Model
• Change the interface between workers and intermediate good firms in the
domestic homogeneous goods sector.
• Labor market in the previous model:
– supply side (household):

Wt
exp (τ t) NtϕCt = .
Pt
– demand side, ith intermediate good firm enters homogeneous labor market
and hires labor required to satisfy demand:
µ ¶−ε
Yt Pi,t
Ni,t = .
At Pt
37
Introducing Labor Market Search and Matching Frictions into Open Economy Model ...
• New setup:
– Each household has a large number of workers, and all are supplied
inelastically to market.
∗ No variation in labor force participation rate. Not a bad approximation,
at least for business cycle variation:
labor force participation rate and HP trend log, real gdp and HP trend
66
9
64
8.5
62
8
60
7.5
1950 1960 1970 1980 1990 2000 1950 1960 1970 1980 1990 2000
correlation, filtered log participation rate filtered log participation rate
with filtered log real gdp = 0.25908 filtered log real gdp
0.02
−0.02
−0.04
1950 1960 1970 1980 1990 2000
38
Homogeneous Final
Domestic Good Consumption
IIntermediate
t di t Foreign
Good Producers Homogeneous
good
Homogeneous
Labor Market
Employment
E l t
Employment Imported Export
Agency
Agency Input Retailers
Foreign
Household
ouse o d Buyers
Household Household
– Each worker is either unemployed and searching, or employed by an

employment agency
∗ the employment agency supplies ς t units of intensity (hours) of labor
services per worker to a competitive labor market
∗ household utility:
∞ workers share equally in household consumption
X z }| {
l−t
Et β { log(Ct+l )
l=0
each household has the same fraction, L, of its workers employed
ς 1+σ L
z}|{
−AL t+l Lt+l }
1 + σL
intensity of labor effort expended by each employed worker ~ς t
– household intertemporal efficiency conditions, (6), (7) unaffected.

– worker search:
∗ intensity of search is fixed and requires no effort or resources
∗ search is ‘undirected’ (i.e., workers do not have useful information about
where the highest wage jobs are).
39
• Employment agencies:
– each agency has a large number of workers.
– 1 − ρ attached workers are randomly selected to separate and go into
unemployment.1
– new workers arrive from unemployment, in proportion to the number of
vacancies posted by the agency in previous period.
– wage paid by employment agency to workers determined.
– intensity of labor effort per worker, ς t, is set according to efficiency
criterion:
workers’ marginal cost of work
z⎡ }| ⎤{
marginal benefit to agency of worker σ ≡ PWAt ≡ Ct
z}|{ ⎢ ALς t L ⎥ scaling
z}|{ z}|{
t t A
σ L z}|{
t
Wt = ⎣ marginal utility of currency ⎦ → w̄t = ALς t ct
z}|{
υt
(36)
– agency or worker unhappy with the match can choose to terminate it at this
point.
1
The assumed constancy of job separation is consistent with the findings reported in Hall (2005b,c) and Shimer
(2005a,b), who report that the job separation rate is relatively acyclical.
40
– agencies are divided into M equal-sized cohorts:
∗ each cohort has a large number of agencies

∗ all agencies in each cohort negotiate the wage with existing workers
(after all current period new arrivals and separations)
∗ wage covers workers at the agency for the duration of the wage contract.
– to characterize what agencies do (bargain with workers and post vacancies),

need:
∗ ‘technology’ for acquiring new workers.

∗ value (surplus) of a worker to the agency, Jt.
∗ equilibrium conditions that characterize vacancy posting decisions.
∗ marginal impact on Jt of a change in the worker wage, Jw,t.
∗ must scale Jt, Jw,t and vacancy posting equilibrium conditions, so
that only variables that appear are those which have a steady state,
independent of initial conditions.
41
– agency in ith cohort posts vacancies, vti, more workers in t + 1

∗ cost of vacancies:
cost parameter that grows with At
z}|{ µ ¶2
κAt Qιtvti
lti
2 lti
lti ~quantity of workers in agency i at t
agency i hiring rate

z}|{
i+1
¡ i 1−ι ¢ i Qιtvti Qtvti
lt+1 = ṽt Qt + ρ lt , ṽti ≡ → ṽtiQ1−ι
t = (37)
lti lti
Qt ~probability that a vacancy is filled
½
1 Gertler-Trigari specification, emphasizes cost of new hires
ι=
0 emphasizes costs of vacancies per se.
∗ costs of posting vacancies rise less rapidly in boom when ι = 1.
42
• Employment Agency Vacancy Posting Problem
– value function of employment agency in the period of contract negotiations

(i = 0), which pays wage rate, ω t :
¡0 ¢
F lt , ω t = present discounted value of profits
⎡ ⎤
vacancy recruiting costs
period t+j revenues per worker z }| {
¡ 0 ¢ M−1 X υ t+j ⎢ z }| { κAt+j i 2⎥
¡ ¢
F lt , ωt = j
β Et max ⎢ (W − ω ) ς − P ṽ ⎥ l j
υ t ṽj ⎣ t ⎦ t+j
t+j t t+j t+j
j=0 t+j
2
⎛ wage rate negotiated at next contract negotiation
⎞
υ t+M ⎝ 0 z }| {
M
+β Et F lt+M , W̃t+M ⎠,
υt
where,
Qιtvti
ṽti ≡ i
lt
43
– Note:
¡0 ¢
F lt , ω t = J (ω t) lt0
J (ωt) ~ present discounted value of profits, per worker

¡0 ¢
F lt , ωt κ ¡ 0¢2
0 ≡ J (ωt) = maxM −1{(Wt − ωt) ς t − PtAt ṽt
lt j
{vt+j } 2
j=0
1
=lt+1 /lt0
υ t+1 h κ ¡ 1 ¢2i ¡z 0 1−ι }| ¢{
+β (Wt+1 − ωt) ς t+1 − Pt+1At+1 ṽt+1 ṽt Qt + ρ
υt 2
2
=lt+2 /lt0
h κ ¡ 2 ¢2i z¡ 0 1−ι ¢}|
¡ 1 ¢{
2 υ t+2 1−ι
+β (Wt+2 − ωt) ς t+2 − Pt+2At+2 ṽt+2 ṽt Qt + ρ ṽt+1Qt+1 + ρ
υt 2
+... +
0
=lt+M /lt0
M υ t+M
¡ ¢ z¡ 0 1−ι ¢¡ 1 1−ι
}|¢ ¡ M−1 1−ι
¢{
+β J W̃t+M ṽt Qt + ρ ṽt+1Qt + ρ · · · ṽt+M−1Qt+M−1 + ρ }.
υt
44
0
– First order
¡ 0 1−ι condition
¢ 1−ι with respect to ṽt (after multiplying the result by
ṽt Qt + ρ /Qt ):
¡ 0 1−ι ¢ 1−ι
0= −PtAtκṽt0 ṽt Qt + ρ /Qt
υ t+1 h κ ¡ 1 ¢2i ¡ 0 1−ι ¢

+β (Wt+1 − Γt,1ωt) ς t+1 − Pt+1At+1 ṽt+1 ṽt Qt + ρ
υt 2
υ h κ ¡ ¢ i¡ ¢¡ 1 ¢
2 t+2 2 2 0 1−ι 1−ι
+β (Wt+2 − Γt,2ωt) ς t+2 − Pt+2At+2 ṽt+2 ṽt Qt + ρ ṽt+1Qt+1 + ρ
υt 2
+...+
M υ t+M
¡ ¢ ¡ 0 1−ι
¢¡ 1 1−ι
¢ ¡ M−1 1−ι
¢
+β J W̃t+M ṽt Qt + ρ ṽt+1Qt + ρ · · · ṽt+M−1Qt+M−1 + ρ }
υt
κ ¡ 0¢2 0
¡ 0 1−ι ¢ 1−ι
= J (ωt) − (Wt − ωt) ς t + PtAt ṽt − PtAtκṽt ṽt Qt + ρ /Qt
2
or,
κ ¡ 0¢2 ρ
J (ωt) = (Wt − ω t) ς t + PtAt ṽt + PtAtκṽt0 1−ι (38)
2 Qt
45
1
– First
¡ 1 order condition
¢ with respect to ṽt+1 (after multiplication by
ṽt+1Q1−ι 1−ι
t+1 + ρ /Qt+1 ):
υ t+1 £ ¡ 1 ¢¤ ¡ 0 1−ι ¢¡ 1 1−ι

¢ 1−ι
β −Pt+1At+1κ ṽt+1 ṽt Qt + ρ ṽt+1Qt+1 + ρ /Qt+1
υt h i¡
2 υ t+2 κ ¡ 2
¢2 0 1−ι
¢¡ 1 1−ι
¢
+β (Wt+2 − ωt) ς t+2 − Pt+2At+2 ṽt+2 ṽt Qt + ρ ṽt+1Qt+1 + ρ
υt 2
+... +
M υ t+M
¡ ¢ ¡ 0 1−ι ¢¡ 1 1−ι
¢ ¡ M−1 1−ι
¢
+β J W̃t+M ṽt Qt + ρ ṽt+1Qt + ρ · · · ṽt+M−1Qt+M−1 + ρ = 0,
υt
which implies (using ṽt0 fonc to substitute out t + 2 and later terms):
υ t+1 £ ¡ 1 ¢¤ ¡ 0 1−ι ¢¡ 1 1−ι

¢ 1−ι
0=β −Pt+1At+1κ ṽt+1 ṽt Qt + ρ ṽt+1Qt+1 + ρ /Qt+1
υt
0
¡ 0 1−ι ¢ 1−ι
PtAtκṽt ṽt Qt + ρ /Qt
υ t+1 h κ ¡ 1 ¢2i ¡ 0 1−ι ¢

−β (Wt+1 − Γt,1ωt) ς t+1 − Pt+1At+1 ṽt+1 ṽt Qt + ρ ,
υt 2
46
¡ 0 1−ι ¢
– divide the last term by ṽt Qt + ρ and rearrange,
" Ã¡ ¢2 !#
1 1
PtAtκ 0 υ t+1 ṽt+1 ṽt+1ρ
ṽt = β (Wt+1 − ωt) ς t+1 + Pt+1At+1κ + 1−ι
Q1−ι
t υ t 2 Qt+1
– to help understand latter first order condition, note:
marginal cost (per lt0 ) of a vacancy

z }|
⎛ ⎞{
2
ṽt0
⎜zQ}|
ι 0⎟
{
t ⎜ tvt ⎟
∂Pt κA
2 ⎝ l0 ⎟
⎜
t ⎠
µ ¶2
Qιt PtAtκ 0 Qt
= PtκAt vt0 = PtκAtQιtṽt0 = 1−ι ṽt 0 .
∂vt0 lt0 Qt lt
period t+1 marginal revenue of a vacancy, per lt0

z }| ¡ ¢{
0 1−ι
∂ (Wt+1 − ω t) ς t+1 ṽt Qt + ρ Qt
= (Wt+1 − ω t) ς t+1 0
∂vt0 lt
47
– continue, for periods t + j, j = 1, 2, ..., M − 2.

³ ´
j j
∗ multiply the first order necessary condition for ṽt+j by ṽt+j Qt+j + ρ /Q1−ι
1−ι
t+j
∗ substitute out for the period t + j + 1 and later terms using the first order
j−1
condition for vt+j−1
– for i = M − 1, substitute out the period t + M term using (38).

– result:
j 1
Pt+j At+j κṽt+j Q1−ι
t+j
µ j+1 2
¶
υt+j+1 ( ṽt+j+1) j+1
ṽt+j+1 ρ
= βEt+j υt+j [(Wt+j+1 − ω t) ς t+j+1 + Pt+j+1At+j+1κ 2 + Q1−ι
],
t+j+1
j = 0, ..., M − 2,
M−1 1
Pt+M−1At+M−1κṽt+M−1 Q1−ι
t+M −1
∙ µ ¶¸
¡ ¢ (ṽt+M )
0 2
0
ṽt+M ρ
= βEt+M−1 υυt+M
t+M
−1
Wt+M − W̃t+M ς t+M + Pt+M At+M κ 2 + Q1−ι t+M
48
– the previous results provide equilibrium expressions for vacancy choices

over time by the cohort of employment agencies that bargain at time t.
– to solve the model, we require equilibrium conditions that hold in the

cross-section of employment agencies at time t:
vacancy decision by agencies that most recently bargained j periods in past, j=0,1,...,M−2
z}|{
Pt At κ
Q1−ι
× ṽtj
t µ j+1 2 ¶
¡ ¢ (ṽt+1 ) j+1
ṽt+1 ρ
= βEt υυt+1t [ Wt+1 − W̃t−j ς t+1 + Pt+1At+1κ 2 + Q1−ι
],
t+1
PtAtκṽtM−1 Q11−ι
t
∙ µ ¶¸
¡ ¢ (ṽt+1)
0 2
0
ṽt+1 ρ
= βEt υυt+1t Wt+1 − W̃t+1 ς t+1 + Pt+1At+1κ 2 + Q1−ι t+1
– next, scale these conditions.
49
– divide by PtAt (use υ t = 1/(PtCt), Ct = Atct): ⎛ ⎞

Ã ! ³ ´2
j+1
κ j ct W̃t−j ṽ
⎜ t+1
j+1
ṽt+1 ρ⎟
1−ι ṽt = βEt c [ w̄t+1 −
P A
ς t+1 + κ ⎝
2
+ 1−ι ⎠],
Qt t+1 t+1 t+1 Qt+1
j = 0, 1, ..., M − 2, where
Wt
w̄t = .
PtAt
– note:
W̃t−j W̃t−j
=
Pt+1At+1 Pt−j At−j π t−j+1 · · · π t+1 exp (∆at−j+1 + · · · + ∆at+1)
¡ ¢ Wt−j
W̃t−j /Wt−j Pt−j At−j
=
π t−j+1 · · · π t+1 exp (∆at−j+1 + · · · + ∆at+1)
wt−j w̄t−j
=
π t−j+1 · · · π t+1 exp (∆at−j+1 + · · · + ∆at+1)
= Gt−j,j+1wt−j w̄t−j
where
inflation and growth factor
z }| { 1 W̃t
Gt−i,i+1 = , wt = .
π t+1 · · · π t−i+1 exp (∆at−j+1 + · · · + ∆at+1) Wt
50
– thus, for j = 0, 1, ..., M − 2,
⎛³ ´2 ⎞
j+1
κ j ct ṽ
⎜ t+1
j+1
ṽt+1 ρ⎟
1−ι ṽt = βEt c [(w̄t+1 − Gt−j,j+1wt−j w̄t−j ) ς t+1 + κ ⎝
2
+ 1−ι ⎠].
Qt t+1 Qt+1
(39)
– for j = M − 1 :
" Ã¡ ¢2
!#
0 0
1 ct ṽt+1 ṽt+1 ρ
κṽtM−1 = βEt (w̄t+1 − wt+1w̄t+1) ς t+1 + κ + 1−ι
Q1−ι
t ct+1 2 Qt+1
(40)
W̃t Wt
wt = , w̄t = .
Wt PtAt
52
• Surplus Enjoyed by Employment Agency With Wage Rate, ω t :
profits earned in employing the work force

z ¡ }| ¢{ value of the firm in case the workers all leave (for unemployment)
z}|{
F lt0, ω t − 0
• Surplus, divided by lt0 :
J (ωt) .
53
• Expressions for scaled surplus and derivative of surplus with respect to wage
j
=0, by optimality of vt+j
M−1 z }| { j
dJ (ωt) ∂J (ω t) X ∂J (ωt) dvt+j
= + j
dωt ∂ωt j=0 ∂vt+j dω t
υ t+1 2 υ t+2
= −ς t − β ς t+1Ωt,1 − β ς t+2Ωt,2
υt υt
2 υ t+M−1
−... − β ς t+M−1Ωt,M−1.
υt
⎧
⎪
⎪ j−1 =growth
Y z
rate of stock of employees
¡ l }| ¢{
employment growth factor
z}|{ ⎨ 1−ι
Ωt,j ≡ ṽt+l Qt+l + ρ j>0 . (41)
⎪
⎪
⎩ l=0 1 j=0
54
– result:
dJ (ωt) υ t+1 2 υ t+2
= −ς t − β ς t+1Ωt,1 − β ς t+2Ωt,2
dωt υt υt
2 υ t+M−1
−... − β ς t+M−1Ωt,M−1,
υt
or,
dJ (ωt) PtAtct PtAtct

≡ Jw,t = −ς t − β ς t+1Ωt,1 − β 2 ς t+2Ωt,2
dωt Pt+1At+1ct+1 Pt+2At+2ct+2
PtAtct
−... − β 2 ς t+M−1Ωt,M−1
Pt+M−1At+M−1ct+M−1

ct z}|{ ct
Jw,t = −ς t − βς t+1 Gt,1 Ωt,1 − β 2ς t+2 Gt,2Ωt,2 (42)
ct+1 ct+2
ct
−... − β M−1ς t+M−1 Gt,M−1Ωt,M−1.
ct+M−1
55
– agency surplus per worker:
κ ¡ 0¢2 0 ρ
J (ωt) = (Wt − ω t) ς t + PtAt ṽt + PtAtκṽt 1−ι
2 Qt
– scale by PtAt:
J (ωt) Wt − ωt κ ¡ 0¢2 ρ
JA,t ≡ = ςt + ṽt + κṽt0 1−ι
PtAt PtAt 2 Qt
or
κ ¡ 0¢2 0 ρ
JA,t = (w̄t − wtw̄t) ς t + ṽ + κṽt 1−ι (43)
2 t Qt
57
• The Worker
– Vti ~value function (in currency units) of a worker attached to an agency in

cohort i, i = 0, .., M − 1, at start of period t, after period t wage has been
determined and all period t separations and job finding have occurred.
– Ut ~value function of a worker (in currency units) at the start of period t,

after all period t separations and job finding have occurred.
58
– value function, i = 0, ..., M − 1 :
disutility of work, converted to currency units by υt

current wage payments z }| {
z }| { 1+σ L
ςt
Vti = W̃t−iς t − AL
(1 + σ L) υ t
value in case of no separation next period
υ t+1 z}|{
i+1
+βEt [ρ × Vt+1
υt
value in case of separation (into unemployment)
z}|{
+ (1 − ρ) × Ut+1 ]
– scale by PtAt :
(44)
i
z }| { ς 1+σ
t
L
ct ct i+1
VA,t = Gt−i,i wt−iw̄t−iς t − AL + βEt [ρVA,t+1 + (1 − ρ) UA,t+1]
1 + σL ct+1
i Vti Ut
VA,t ≡ , UA,t ≡ .
PtAt PtAt
59
– for Nash bargaining, it is useful to have the derivative of Vt0 with respect to
the wage rate negotiated at the start of period t by workers in agencies of
cohort i = 0.
– denote this derivative, when evaluated at an arbitrary wage rate, ω t, by

0
Vwt (ω t) :
M−1
X M−1
X
j υ t+j PtAtct
Vw0 (ω t) = (βρ) Etς t+j = (βρ)j Etς t+j
j=0
υt j=0
Pt+j At+j ct+j
M−1 growth and inflation factor
X j z}|{ ct
= (βρ) Etς t+j Gt,j . (45)
j=0
ct+j
– note: effort, ς t, is independent of ω t. It is determined by the efficiency

criterion.
60
– value function of unemployed worker, Ut :
unemployment compensation
z }| {u
Ut = PtAtb
prob of finding a job value function of worker that finds a job, before knowing what agency the job is in
υ t+1 z}|{ z}|{x
+βEt [ ft × Vt+1
υt
+ (1 − ft) Ut+1],
probability that a worker who finds a job, ends up in an agency that was in cohort i in t−1
z }| {
hiring rate
M−1
z }| {
X i
ṽt−1 Q1−ι l i
t−1 t−1
Vtx = × Vti+1 (46)
i=0
mt−1
job matches made by agencies in cohort j
total number of matches M−1
X z }| {
z}|{
mt = ṽtj Q1−ι
t lt
j
(47)
j=0
61
– scaling by PtAt :
u ct x
UA,t = b + βEt [ftVA,t+1 + (1 − ft) UA,t+1] (48)
ct+1
x Vtx Ut
VA,t ≡ , UA,t ≡ .
PtAt PtAt
– other labor market relations:
total hours worked M−1
z}|{ X
Nt = ςt ltj , (49)
j=0
total employment M−1
z}|{ X
Lt = ltj , (50)
j=0
worker job finding rate
z}|{ mt
ft = , (51)
1 − Lt
matching function
z }| { total vacancies
z}|{
M−1
X j mt
σ 1−σ
mt = σ m (1 − Lt) vt , vt = vt , Qt =
j=0
vt
63
• Nash bargaining problem

⎛ 0
⎞η
surplus of a representative worker who bargains with firm on behalf of all workers, lt
⎜value function of a worker, who receives wage, ωt }|
z {⎟
⎜ z }| { the alternative to a successful bargain is unemployment⎟
z}|{
⎜ 0 ⎟
max ⎜ Vt (ω t) − Ut ⎟
ωt ⎜ ⎟
⎝ ⎠
⎛ ⎞(1−η)
surplus of firm, divided by lt0
z }| {
×⎝ J (ωt) ⎠
– first order necessary condition for solution to Nash bargaining problem

(after scaling):
0
Vw,t Jw,t
η 0 + (1 − η) = 0, evaluated at ω t = W̃t. (52)
VA,t − UA,t JA,t
64
• changes to previous version of the model:

– clearing in homogeneous goods -
imports used in production of final consumption goods

z }| η {
p∗t Nt = gt + (1 − ωc) (pct) c ct (53)
resources used up in producing vacancies

exports
z }| {
z }| { κ X ³ j ´2 j
M−1
+(pxt)−ηf yt∗ + ṽt lt .
2 j=0
– price setting equation, (4) replaced by:
= intermediate good firm marginal cost

ε z }| {
Kt = (1 − ν t) w̄t (1 − ψ + ψRt) + Etβθπ̄ εt+1Kt+1 (54)
ε−1
65
• extra equations and variables that have been introduced because of the search
and matching:
variable equation(s)
w̄t (36)
lti (37)
ṽti (39) , (40)
Jw,t (42)
JA,t (43)
i
VA,t (44)
0
Vw,t (45)
x
VA,t (46)
mt (47)
UA,t (48)
Qt (52)
ςt (49)
Lt (50)
ft (51)
66
Financial Frictions
• Financial frictions occur when the people who implement projects are different
from the people who have the surplus funds needed to finance those projects.
– In practice, simple frictionless sharing rules are not feasible.
• Asymmetric information creates incentive problems in relationship between

borrowers and lenders:
– borrower might report ‘bad outcomes’, and pocket the difference between
actual and reported outcomes.
∗ borrowers might put in little effort, and then claim ex post that ‘bad
outcomes’ were due to bad luck.
∗ borrowers may undertake excessively risky investments, in case losses
are bounded below and potential gains not bounded above.
67
Financial Frictions ...
• Arrangements between borrowers and lenders:
– can be a source of shocks to the economy
– can alter the propagation of standard shocks (e.g., government spending,

technology, ...)
∗ example: nominal interest is typically non-state contingent, so when

price moves unexpectedly wealth is transferred between borrowers and
lenders (Irving Fisher ‘debt deflation’). Has real effects when these
people are very different.
• We will review the financial frictions suggested by Bernanke, Gertler, Gilchrist

(1999), as implemented by Christiano, Motto and Rostagno (2003, 2007),
Christiano, Trabant and Walentin (2008).
• Other notes on financial frictions,

see http://faculty.wcas.northwestern.edu/~lchrist/d16/d1608/syllabus.htm
68
• We will stress financial frictions associated with the accumulation and

management of the society’s physical stock of capital (this involves long-term
finance which surely is where to biggest frictions exist!)
• In RBC model, the household does the saving and also the investing. That
model abstracts completely from financial frictions.
• We will assume that households save, but they do not own or manage the
allocation of capital.
• First, we must describe the way capital is used in the model, and the technology
for accumulating capital.
69
Homogeneous Final
Domestic Good Consumption Foreign
Homogeneous
good
Exporters Foreign
Buyers
Intermediate Homogeneous Capital

Good Producers Market
Entrepreneurs
Labor Market Capital Producer

Banks
Households
At end of period t:
Homogeneous 1. Entrepreneurs sell old capital to capital producers
Domestic Good and pay off debt.
2 Households lend funds to banks
2. banks.
3. Banks make loans to entrepreneurs, depending on
quantity of entrepreneurial net worth.
4. Entrepreneurs
p buyy new capital
p from capital
p
producers.
5. In t+1 entrepreneurs supply capital to rental market.
6. At end of t+1, go back to 1.
Intermediate Homogeneous Capital

Good Producers Market
Entrepreneurs
Capital Producer
Banks
Households
• Intermediate good production function:
– Previous specification:
Yi,t = AtNi,t.
– Now:
Yi,t = (AtNi,t)1−α Ki,t

α
, 0 < α < 1.
– Capital used by ith intermediate good producer, aggregate capital:

Z 1
Ki,t, Kt = Ki,tdi.
0
– To avoid conflict with notation in price equation, we now replace Kt in with

Ktp.
70
• Optimization by intermediate good firms.
– price-setting problem completely unchanged.
∗ marginal cost still constant.
∗ price-setting efficiency conditions, ((2), (3)), and (4) becomes
ε
Ktp = p
mct + Etβθπ̄εt+1Kt+1 , (55)
ε−1
where mct denotes marginal cost in domestic currency units/Pt (see
below for more).
71
– Production efficiency by intermediate good firms.
– firm enters in competitive factor markets and minimizes cost:

£ ¤ h i
min τ dt (1 − ν t) Wt (1 − ψ + ψRt) Ni,t + Ptrtk Ki,t +λt Yi,t − (AtNi,t)1−α Ki,t
α
Ni,t ,Ki,t
where
rtk rental rate of capital, in units of domestic homogeneous good

τ dt shock to marginal cost that does not affect production function
λt multiplier, nominal marginal cost
νt labor market subsidy
– foncs:
Wt [1 − ψ + ψRt] = λt (1 − α) (At)1−α (Ni,t)−α Ki,t
α
Ptrtk = λtα (AtNi,t)1−α Ki,t

α−1
72
– rewriting two foncs:

¡ ¢ ¡ ¢ ¡ k ¢α
1 1−α 1 α
λt τ dt 1−α α − ψ + ψRt])1−α
rt Pt (Wt [1
mct ≡ =
Pt At1−αPt
µ ¶1−α µ ¶α
1 1 ¡ k ¢α
d
= τt rt (w̄t [1 − ψ + ψRt])1−α (56)
1−α α
Wt
w̄t = ,
PtAt
– also:
1 Wt [1 − ψ + ψRt]
mct = τ dt
Pt (1 − α) (At)1−α (Ni,t)−α Ki,t
α
w̄t [1 − ψ + ψRt]
= τ dt ³ ´α (57)
kt
(1 − α) exp(∆a t )Nt
73
• Domestic homogeneous good output: µ ¶ε

Pt∗
Yt = p∗t (AtNt)1−α Ktα, p∗t ≡
Pt
∙Z ¸ −1ε h i −1ε
1 ¡ ¢−ε
Pt∗ ≡ −ε
Pi,t di = (1 − θ) P̃t−ε + θ Pt−1
∗
0
∙Z 1 ¸ 1−ε
1
h i1
(1−ε) (1−ε) (1−ε) 1−ε
Pt ≡ Pi,t di = (1 − θ) P̃t + θPt−1
0
– p∗t has law of motion, (1).
– Scale Kt+1 by At :
µ ¶α
kt Yt Kt At
yt = p∗t (Nt)1−α , yt ≡ , kt ≡ , exp (∆at) = .
exp (∆at) At At−1 At−1
(58)
74
• Capital accumulation technology:

µ µ ¶¶
It
Kt+1 = (1 − δ) Kt + 1 − S̃ It.
It−1
S̃ 00
S̃ (x) = (x − 1)2 , S̃ (1) = S̃ 0 (1) = 0.
2
• ‘Cost of change in investment’ specification:
– Empirical motivations:
∗ Christiano, Eichenbaum and Evans (JPE, 2005): helps account for
hump-shaped response of investment to monetary policy shocks.
∗ Sherwin Rose, JPE, empirical evidence from housing construction.
∗ Matsuyama, 1984, ‘A Learning Effect of Model of Investment: An
Alternative Interpretation of Tobin’s Q’, manuscript.
∗ Conventional (‘static’) adjustment costs rejected by data (more below).
– Economic motivations:
∗ ‘learning-by-doing’ Matsuyama (1984)
∗ D. Lucca, 2008, ‘Resuscitating Time-to-Build’, manuscript, US BOG.
76
• Capital production is operated by identical, competitive ‘capital producers’
– buy investment goods, It, for price Pt

– buy ‘old capital’, x, for price PtPk0,t
– sell ‘new capital’, y, for price PtPk0t
– no finance required (simplification)
– technology for building capital:
denote price by Pt Pk0 t

z}|{
Kt+1 = (1 − δ) Kt
µ µ ¶¶ transformed one-for-one from homogeneous output, so price = Pt

It z}|{
+ 1 − S̃ It
It−1
– Divide by At : µ µ ¶¶
1−δ exp (∆at) it It
kt+1 = kt + 1 − S̃ it, it ≡ . (59)
exp (∆at) it−1 At
77
• Objective:
marginal value (to household) of unit of currency

∞
X z }| {
max Et β j υ t+j
{It+j ,xt+j }
j=0
profits in currency units
z½ ∙ µ µ }|
¶¶ ¸ ¾{
It+j
× Pt+j Pk0,t+j xt+j + 1 − S̃ It+j − Pt+j Pk0,t+j xt+j − Pt+j It+j .
It+j−1
– Demand for x is indeterminate, so in equilibrium it is simply equal to

supply:
xt = (1 − δ) Kt.
78
– First order necessary condition for optimization of It (after rearranging):
static marginal cost of capital = price of investment ÷ marginal product of investment in producing new capital
z }| {
1
Pk0,t = ³ ´ ³ ´
It It It
1 − S̃ It−1 − S̃ 0 It−1 It−1
1 υ t+1Pt+1
− ³ ´ ³ ´ βEt
1 − S̃ It
− S̃ 0 It It υ tPt
It−1 It−1 It−1
>0 if planning It+1 /It >1, so that It ↑ saves on t+1 adjustment costs
z µ }| ¶ µ ¶{2
It+1 It+1
× Pk0,t+1S̃ 0
It It
79
– Compare to implication of ‘conventional’ adjustment costs:
µ ¶
It
Kt+1 = (1 − δ) Kt + κ Kt, κ0 > 0, κ00 < 0.
Kt
1
P k 0 ,t = ³ ´
It
κ0 Kt
– Conventional theory rejected because variables other than It enter.
– Adjustment cost in change of investment predicts this result.
– Connection of present theory to current literature needs additional explo-

ration.
80
• Scaled equations associated with capital producer:
1
Pk0,t = ³ ´ ³ ´
it exp(∆at ) it exp(∆at ) it exp(∆at )
1 − S̃ it−1 − S̃ 0 it−1 it−1
1
− ³ ´ ³ ´ (60)
it exp(∆at ) it exp(∆at ) it exp(∆at )
1 − S̃ it−1 − S̃ 0 it−1 it−1
µ ¶µ ¶2
ct it+1 exp (∆at+1) it+1 exp (∆at+1)
× βEt Pk0,t+1S̃ 0
exp (∆at+1) ct+1 it it
81
• Entrepreneurs
– the only people who can own and allocate (rent out) the economy’s capital
stock.
– find it profitable to own more capital than they can afford (they borrow from
banks).
– asymmetric information creates financial frictions.
• The entrepreneur in t and t + 1 :

– Period t :
∗ State of entrepreneur: level of net worth, Nt+1
∗ Entrepreneur with net worth, N (‘N−type entrepreneur’) borrows:
purchase of capital by N−type entrepreneur

z }| {
N N
Bt+1 = PtPk0,tKt+1 − Nt+1.
82
∗ After purchasing capital, N−type entrepreneur receives idiosyncratic

shock:
N N
Kt+1 → Kt+1 ω, ω ~iid cdf Ft (ω)
ω lognormal across entrepreneurs with Eω = 1, variance(log ω) = σ 2t .
∗ Entrepreneur sees ω, bank must pay a ‘monitoring cost’ to see it. Profit
sharing too expensive for banks.
∗ Good arrangement:
· Bank gives N−type entrepreneur ‘standard debt contract’, which

N N
specifies a level of borrowing, Bt+1 , and an interest rate, Zt+1 .
· The bank only monitors the (bankrupt) entrepreneurs who say they
cannot pay the stipulated interest rate.
84
– Period t + 1 :
∗ entrepreneur who purchased one unit of capital in t earns:
z rent
}| { zsale of undepreciated
}|
capital
{ average return on capital
k
Pt+1rt+1ω + (1 − δ)Pt+1Pk0,t+1ω z}|{
k
≡ Rt+1 × ω.
PtPk0,t
∗ resources available to N−type entrepreneur with shock, ω, for paying
back debt:
k N
PtPk0,tRt+1 ωKt+1
∗ cutoff idiosyncratic shock, ω̄N t+1:
N
N 1 Bt+1
ω̄t+1 = Zt+1 k N
Rt+1 PtPk0,tKt+1
ω < ω̄ N
t −→ N − type entrepreneur declares bankruptcy,
is monitored, and hands over everything to bank
k N
(monitoring cost, μRt+1 ωPtPk0,tKt+1 , is paid by bank)
ω ≥ ω̄ N
t → pays interest to bank, and is not bankrupt.
85
• Banks
– zero profit condition for each type of entrepreneur:
net earnings from bankrupt entrepreneurs

earnings from non-bankrupt entrepreneurs z Z }| {
z£ ¡ N}|¢¤ N N { ω̄N
t+1
k N
1 − Ft ω̄ t+1 Zt+1Bt+1 + (1 − μ) ωdFt (ω) Rt+1 PtPk0,tKt+1
0
money owed by banks to households, not contingent on t+1 shocks (Fisher deflation)
z }| {
N
= RtBt+1 .
– the only source of funds available for banks to repay debt in t + 1 is assumed
to be their receipts from entrepreneurs.
87
N N
– substitute out for Zt+1 Bt+1 in zero profit condition:
( Z )
£ ¡ ¢¤ ω̄N
t+1
N
RtBt+1 = 1 − Ft ω̄N
t+1 ω̄N
t+1 + (1 − μ)
k
ωdFt (ω) Rt+1 N
PtPk0,tKt+1
0
⎡share of entrepreneurial earnings received by banks share of entrepreneurial earnings used up in monitoring⎤
z }| { z }| {
= ⎣ N
Γt(ω̄t+1) − μGt(ω̄N t+1)
⎦
entrepreneurial earnings
z }| {
k N
× Rt+1PtPk0,tKt+1
Z ω̄N
t+1 £ ¤
Gt(ω̄N
t+1) ≡ ωdFt(ω), Γt(ω̄ N
t+1 ) ≡ ω̄N
t+1 1 − Ft(ω̄t+1) + Gt(ω̄N
N
t+1 )
0
– can show: in neighborhood of equilibrium, object in square brackets is

increasing in ω̄ N
t+1
k
⇒model has (sensible) implication that shock which drives up Rt+1 ,
drives down bankruptcy rate
88
– Zero profit condition, divided by RtNt+1:
assets to net worth ratio

z }| {
N
£ N N
¤ k
Rt+1 N N
N
PtPk0,tKt+1
t − 1 = Γt(ω̄t+1) − μGt(ω̄t+1) , ≡
Rt t t
Nt+1
– two parameters of the debt contract:
N N
∗ loan amount, Bt+1 , and interest rate, Zt+1
∗ or, equivalently, ω̄N
t+1 and t
N
– in equilibrium, loan contract chosen to maximize (ex ante) entrepreneurial

utility (scaled by entrepreneur’s opportunity cost):
R∞ £ k N N N
¤ ½ ¾
Et ω̄t+1 Rt+1ωPtPk0,tKt+1 − Zt+1Bt+1 dFt(ω) £ ¤ R k
t+1
= Et 1 − Γt(ω̄ N
t+1 )
N
t ,
RtNt+1 Rt
subject to bank zero profit condition.
89
– Lagrangian representation of problem:

£ N
¤ k N
1 − Γt(ω̄ t+1) Rt+1 t
max Et{
t { t+1 }
N , ω̄ N Rt
state-by-state zero profit condition
z£
Ã }| ¤ !{
N N k N
Γt (ω̄ t+1 ) − μGt (ω̄ t+1 ) Rt+1 t
+λN
t+1 − N
t + 1 }.
Rt
– First order necessary conditions for optimality:

½ µ ¶¾
£ ¤ k
Rt+1 N
£ ¤ k
Rt+1
Et 1 − Γt(ω̄N
t+1 ) + λt+1 Γ (ω̄
t t+1
N
) − μG (ω̄
t t+1
N
) −1 = 0
Rt Rt
0 N N
£ 0 N 0 N
¤
−Γt(ω̄ t+1) + λt+1 Γt(ω̄ t+1) − μGt(ω̄ t+1) = 0
£ ¤ R k
t+1
Γt(ω̄ N N
t+1 ) − μGt (ω̄ t+1 )
N N
t − t + 1 = 0.
Rt
90
– Substituting out for λN

t+1:
£ N N
¤ Rtk N N
Γt−1(ω̄ t ) − μGt−1(ω̄t ) t−1 − t−1 + 1 = 0
Rt−1
⎧ ³£ ¤ ´⎫
⎪
⎨£ Γ0
(ω̄ N
) Γ (ω̄ N
) − μG (ω̄ N
)
k
Rt+1
− 1 ⎪
⎬
¤ R k
t+1 t t+1 t t+1 t t+1 R t
Et 1 − Γt(ω̄ N t+1 ) + 0 N 0 N
= 0.
⎪
⎩ Rt Γt(ω̄t+1) − μGt(ω̄t+1) ⎪
⎭
– key properties of solution:
∗ t, ω̄ t independent of N (drop superscript, N )
∗ all entrepreneurs pay same interest rate, and N−type entrepreneur

receives loans in proportion to his/her net worth.
t k
Bt+1 = Nt+1 ( t − 1) , Zt+1 = Rt+1 ω̄t+1.
t −1
91
• One more equilibrium condition: law of motion for aggregate net worth.
– density of entrepreneurs with net worth N , ft (N) , aggregate net worth,

N̄t+1 Z
N̄t+1 = Nft (N) dN.
– N−type entrepreneurs have, after paying back bank loans,
VtN = [1 − Γt−1(ω̄ t)] Rtk Pt−1Pk0,t−1KtN
– multiply by f (N) and integate over all N, to obtain:
Vt = [1 − Γt−1(ω̄ t)] Rtk Pt−1Pk0,t−1Kt,

where
Z Z
Vt = VtN ft (N) dN, Kt = KtN ft (N) dN.
N N
92
– alternative representation of Vt
t−1 t=1−Γ (ω̄ )

z
½ }| Z ¾{
ω̄t
Vt = 1 − ω̄t [1 − Ft−1(ω̄ t)] − ωdFt−1(ω) Rtk Pt−1Pk0,t−1Kt
0
= Rtk Pt−1Pk0,t−1Kt
this represents earnings of banks, which must equal Bt Rt−1 =Rt−1 (Pt−1 P
k ,t−1 0 Kt −N̄t )
zµ }|
Z ω̄t ¶ {
− ω̄t [1 − Ft−1(ω̄t)] + (1 − μ) ωdFt−1(ω) Rtk Pt−1Pk0,t−1Kt
0
Z ω̄ t
−μ ωdFt−1(ω)Rtk Pt−1Pk0,t−1Kt
0
gross interest paid by entrepreneurs as a whole on their loans
z" R ω̄t }| #{
μ ωdFt−1(ω)Rtk Pt−1Pk0,t−1Kt
= Rtk Pt−1Pk0,t−1Kt − Rt−1 + 0
Bt.
Pt−1Pk0,t−1Kt − N̄t
93
– after Vt is determined, entrepreneur exits with probability 1 − γ t and

survives with probability γ t, 1 − γ t new entrepreneurs ‘born’ without any
wealth.
– all entrepreneurs who pass into the next period receive a transfer, Wte
– law of motion of aggregate net worth:
N̄t+1 = γ t [1 − Γt−1(ω̄ t)] Rtk Pt−1Pk0,t−1Kt + Wte.
– scaling by PtAt :
Pk0,t−1kt
nt+1 = γ t [1 − Γt−1(ω̄ t)] Rtk + wte
π t exp (∆at)
N̄t+1 e Wte
nt+1 ≡ , w ≡ .
PtAt t PtAt
94
• Summary of three equilibrium conditions associated with entrepreneur:
– aggregate, scaled, entrepreneurial net worth:
Pk0,t−1kt
nt+1 = γ t [1 − Γt−1(ω̄t)] Rtk + wte (61)
π t exp (∆at)
– zero profit condition for banks:
Rtk
[Γt−1(ω̄t) − μGt−1(ω̄t)] t−1 − t−1 +1=0 (62)
Rt−1
– optimality condition for entrepreneurial contract
⎧ ³ ´⎫
⎪
⎨ k Γ0
(ω̄ ) [Γ (ω̄ ) − μG (ω̄ )]
k
Rt+1
− 1 ⎪
⎬
Rt+1 t t+1 t t+1 t t+1 Rt
Et [1 − Γt(ω̄ t+1)] + 0 (ω̄ 0 (ω̄ = 0.
⎪
⎩ R t Γt t+1 ) − μGt t+1 ) ⎪
⎭
(63)
95
• Clearing of domestic homogeneous goods.
– modification to (14):
yt = (1 − ωc) (pct)ηc ct + xt + gt + it + dt, (64)
where yt is defined in (58) and scaled monitoring costs, dt, are:
μGt−1(ω̄t)Rtk Pt−1Pk0,t−1Kt μGt−1(ω̄ t)Rtk Pk0,t−1kt

dt = =
PtAt π t exp (∆at)
– net worth of exiting entrepreneurs:
(1 − γ t) Vt.
∗ a fraction, 1 − Θ, of (1 − γ t) Vt is taxed and transferred lump sum to

households, and Θ is consumed by entrepreneurs. We suppose Θ is so
small that consumption by entrepreneurs can be ignored.
97
– Concluding observations on financial frictions
– New impulses:
∗ Shock up in γ t+1 : raises amount of wealth in hands of entrepre-

neurs....model’s way of capturing a ‘bubble’ jump in the stock market
(i.e., one not obviously linked to fundamentals)
∗ Shock up in σ t : increases risk associated with entrepreneurial lending -

resembles current subprime lending crisis.
‘don’t know who is sitting on the bad mortgage loans and who is not’
98
– New propagation:
∗ shock that drives down rental earnings or price of capital reduces net
worth and restricts ability of entrepreneurs to borrow.
∗ entrepreneurs’ reduced ability to buy capital causes them to reduce

purchases of goods by capital producers and induces a fall in output and
employment
∗ new friction: rigidity in nominal rate of interest affects propagation

mechanism
· acts as an accelerator for expansionary shocks that raise the price level
· acts as a moderating factor for expansionary shocks that reduce the
price level.
99
• Collecting the equations of the model, following 15:

– price optimization:
p ε p
Kt = mct + Etβθπ̄ εt+1Kt+1 , (65)
ε−1
Ftp = 1 + Etπ̄ ε−1
t+1 βθFt+1
p
(66)
∙ ε−1 ¸ 1−ε
1
1 − θπ̄ t
Ktp = Ftp (67)
1−θ
– investment efficiency condition:
∙ µ ¶ µ ¶ ¸
it exp (∆at) it exp (∆at) it exp (∆at)
1 − S̃ − S̃ 0 Pk0,t
it−1 it−1 it−1
(68)
µ ¶µ ¶2
ct 0 it+1 exp (∆at+1) it+1 exp (∆at+1)
= 1 − βEt Pk ,t+1S̃
0
exp (∆at+1) ct+1 it it
– household labor optimization:
exp (τ t) Ntϕct = w̄t (69)
100
– average rate of return on capital:
rtk + (1 − δ)Pk0,t
Rtk = π̄ t (70)
Pk0,t−1
– cost minimization:
µ ¶1−α µ ¶α
1 1 ¡ k ¢α
d
mct = (1 − ν t) τ t rt (w̄t [1 − ψ + ψRt])1−α (71)
1−α α
d w̄t [1 − ψ + ψRt ]
mct = (1 − ν t) τ t ³ ´α (72)
kt
(1 − α) exp(∆at)Nt
– domestic homogeneous output

µ ¶α
kt
yt = p∗t (Nt)1−α (73)
exp (∆at)
μGt−1(ω̄t)Rtk Pk0,t−1kt
yt = (1 − ωc) (pct)ηc ct + xt + gt + it + (74)
π̄ t exp (∆at)
101
– current account µ ¶ηc f f

pct st R Φt−1 a
aft + pm,c
t ωc ct = pctqtpxtxt + t−1 t−1
(75)
pm,c
t π̄ t exp (∆at)
– capital law of motion:
µ µ ¶¶
1−δ exp (∆at) it
kt+1 = kt + 1 − S̃ it (76)
exp (∆at) it−1
– conditions related to entrepreneurs:
Pk0,t−1kt
nt+1 = γ t [1 − Γt−1(ω̄t)] Rtk + wte (77)
π̄ t exp (∆at)
½ ¾
Rtk Pk0,t−1kt
0 = [Γt−1(ω̄t) − μGt−1(ω̄ t)] −1 +1 (78)
⎧ Rt−1 n t
³ ´⎫
⎪
⎨ k Γ0
(ω̄ ) [Γ (ω̄ ) − μG (ω̄ )]
k
Rt+1
− 1 ⎪
⎬
Rt+1 t t+1 t t+1 t t+1 Rt
0 = Et [1 − Γt(ω̄ t+1)] + 0 (ω̄ 0 (ω̄ (79)
⎪
⎩ R t Γt t+1 ) − μGt t+1 ) ⎪
⎭
102
– 10 remaining equations:
monetary policy - (5); household intertemporal first order conditions - (6),

(7); price equations - (8), (9), (10), (13), (1); demand for exports - (11);
exporter equilibrium condition - (12).
– 25 variables:
st, ct, πct, pct, pm,c
t , qt , π̄ t , px f
t , at , xt , nt+1 , R k
t , Pk0 ,t ,
kt+1, ω̄ t, Nt, it, p∗t , Ktp, Ftp, mct, rtk , w̄t, yt, Rt
– shocks:
τ t, εR,t, Rtf , ytf , πft , ∆at, σ 2t , γ t, gt, ytf , φ̃t, τ dt.
103
• Steady state of the model

– as before:
g = η g y, ā = ηay,
where ā is the parameter controlling Φt (see (16)).
– algorithm for solving for the steady state:

∗ π̄ c, π̄, R, p∗, F p can be computed from (20), (21), (18), (22), (23).
∗ from (65),
ε
ε−1 mc
Kp =
1 − βθπ̄ ε
∗ combining this with (67),
∙ ε−1
¸ 1−ε
1
ε
∙ ε−1
¸ 1−ε
1
ε−1 p 1 − θπ̄ ε − 1 1 − βθπ̄ 1 − θπ̄
mc = F [1 − βθπ̄ ε] =
ε 1−θ ε 1 − βθπ̄ ε−1 1−θ
104
∗ (18), (19) imply
Rf R
=
πf πc
which can be used to solve for Rf given π f .
∗ steady state depreciation, s, can be computed from the inflation

differential:
qt → q implies (see (13)) sπ f = πc.

∗ (17), (18) imply
sRf Φ = R,
or after multiplication by π f and rearranging,
Rf R f
Φ = , so (see (32)) Φ = 1 and at = ā (see (16))
πf πc
105
– equation (68), together with the facts, S = S 0 = 0 in steady state implies
Pk0 = 1.
– rest of the algorithm solves a single non-linear equation, (11), in a single

unknown, ϕ̃ = pcq .
– set
ϕ̃ = pcq.
– use (27), (28), (26):
pm,c = ϕ̃Rν,∗
x
R
px = ,
ϕ̃
h i 1−η1
1−η
pc = (1 − ωc) + ω c (pm,c) c
c
ϕ̃
q = c.
p
106
– consider the following 11 unknowns:
n, c, w̄, k, N, y, x, i, ω̄, rk , Rk
in the following 11 equations:
k
¡ k ¢
(1)R = π̄ r + 1 − δ
µ ¶α
k
(2)y = p∗ (N)1−α
∙ exp¸ (∆a)
1−δ
(3)i = 1− k
exp (∆a)
(4)w̄ = exp (τ ) N ϕc
³ c ´ηc
f m,c p sRf af
a + p ω c pm,c c − π̄ exp(∆a)
(5)x =
pcqpx
k k
(6)n = γ [1 − Γ(ω̄)] R + we,
π̄ exp (∆a)
107
µ ¶1−α µ ¶α
1 1 ¡ k ¢α
(7)mc = (1 − ν) τ d
r (w̄ [1 − ψ + ψR])1−α
1−α α
w̄ [1 − ψ + ψR]
(8)mc = (1 − ν) τ d ³ ´α
k
(1 − α) exp(∆a)N
c ηc μG(ω̄)Rk k
(1 − ωc) (p ) c + x + i + π̄ exp(∆a)
(9)y =
1 − ηg
µ k
¶
R k
(10)0 = [Γ(ω̄) − μG(ω̄)] −1 +1
R ³ n ´
k
0 R
Rk Γ (ω̄) [Γ(ω̄) − μG(ω̄)] R − 1
(11)0 = [1 − Γ(ω̄)] +
R Γ0(ω̄) − μG0(ω̄)
– to solve these 11 equations, fix rk . Compute Rk using (1). Solve for ω̄ using
(11). Solve for k/n using (10). Solve for n, k by rewriting (6):
we
n= k/n
.
k
1 − γ [1 − Γ(ω̄)] R π̄ exp(∆a)
108
Solve for i using (3). Solve for w̄ by rewriting (7):

" # 1−α
1
mc
w̄ = ¡ 1 ¢1−α ¡ 1 ¢α α 1−α
.
(1 − ν) τ d 1−α α (rk ) (1 − ψ + ψR)
Solve for N by rewriting (8):
∙ ¸− α1
k d w̄ [1 − ψ + ψR]
N= (1 − ν) τ
exp (∆a) (1 − α) mc
Solve for y using (2). Substitute out for x in (9) using (5). Solve the
resulting version of (9) for c :
¡ ¢ af − π̄ sR
f af
exp(∆a) μG(ω̄)Rk k
1 − η g y + pcqpx − i − π̄ exp(∆a)
c= pc ηc
pm,c ω
( )
(1 − ω c) (pc)ηc +
c pm,c
pc qpx
Adjust rk until (4) is satisfied.
109
– adjust ϕ̃ until the demand for exports, (11), is satisfied:
x = (px)−ηf y f .
– alternatively, one could simply fix ϕ̃ and let the previous equation define y f .
– we evaluate G (ω̄) and Γ (ω̄) .
∗ the lognormal distribution has two parameters, a mean and a variance,

σ 2x.
∗ the mean is determined by the requirement, Eω = 1
∗ the variance, σ 2x, is determined by F (ω̄) , which we fix a priori.
for a given ω̄, we solve for"σ x using #
1 2
log (ω̄) + 2 σ x
F (ω̄) = prob v <
σx
110
∗ given ω̄ and σ x, we solve "for G (ω̄) , Γ (ω̄) using #

log (ω̄) + 12 σ 2x
G(ω̄) = prob v < − σx
σx
Γ (ω̄) = ω̄ [1 − F (ω̄)] + G(ω̄).
∗ also,
G0(ω̄) = ω̄F 0(ω̄)

Γ0(ω̄) = 1 − F (ω̄) − ω̄F 0(ω̄) + G0(ω̄)
= 1 − F (ω̄) > 0,
so that
Γ0(ω̄)
Γ0(ω̄) − μG0(ω̄)
1 − F (ω̄)
=
1 − F (ω̄) − μω̄F 0(ω̄)
111

openeconomy

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

openeconomy

Uploaded by

Copyright:

Available Formats

...

Extensions on the Basic Model

• Open economy (based on work of Adolfson-Laséen-LindéVillani and

• Search and matching in labor market (based on work of Gertler-Sala-Trigari

• Financial frictions (based on work of Bernanke-Gertler-Gilchrist and

– Equilibrium conditions associated with price setting:

– computing the steady state

– summary of the endogenous and exogenous variables of the model, as well

– several computational experiments to illustrate the properties of the model.

• Modifications to basic model to create open economy

Labor Market Foreign

• Monetary policy: three approaches

• Household budget constraint

• optimality of household foreign asset decision (verify this by solving

• Optimality of household domestic bond decision:

• Final domestic consumption goods, Ct

Ctd ~one-for-one transformation on domestic homogeneous output good, price Pt

– Ctm is produced by competitive firm, which converts foreign homogeneous

∗ Setting price equal to marginal cost:

– Consumption good inflation:

π ct ≡ c = = π̄ t ¡ m,c ¢1−ηc . (10)

• Clearing in domestic homogeneous goods market:

output of domestic homogeneous good, Yt

goods used in production of final consumption, Ct

= (1 − ωc) (pct)ηc Ct + Xt + Gt.

Atp∗t Nt = (1 − ωc) (pct)ηc Ct + Xt + Gt,

currency flowing into the country

receipts from exportst = StPtxXt.

– Divide current account by PtAt :

StAft+1 StPtf ³ ´ µ pc ¶ηc S P x S R f

where aft is ‘scaled real, domestic value of foreign assets’:

– Captures, informally, the possibility that there is a shock to the required

– φ̃s is an important reduced form feature, designed to correct an important

– household intertemporal efficiency conditions:

– other steady state conditions:

pc, pm,c, q, px, c, x, af , R, Φ, π̄, K, F, p∗, πc, s

g = ηg y, ā = η ay, where y = p∗N.

– algorithm for solving for the steady state:

– rest of the algorithm solves a single non-linear equation in a single unknown.

which can be solved linearly for x.

– evaluate (31) and adjust ϕ̃ until it is satisfied. In practice, we set ϕ̃ = 1 and

• Uncovered interest rate parity puzzle and Φbt

– subtract equations (17)

– totally differentiate the object in square brackets, and evaluate in steady

rt − log (R) = log St+1 − log St − log s + rtf − log Rf + Φ̂t.

– Uncovered Interest Parity (UIP)

rt > rtf → there must be an anticipated depreciation (instantaneous appreciation) of the

· β̂ ' −.75, so UIP rejected (that’s the UIP puzzle)

∗ VAR impulse responses by Eichenbaum and Evans (QJE, 1992)

· UIP puzzle: rt ↑ and expected appreciation of the currency represents

that expectations of depreciation emerge. In this way, UIP leads to

· model’s resolution of the UIP puzzle: when rt ↑ the return required

– 16 endogenous variables: pct, pm,c

– exogenous variables: Rtf , ytf , φ̃t, gt, εR,t, ∆at, τ t, π ft .

– the model was solved in the manner described above:

• iid shock, 0.01, to εR,t.

We now consider a monetary policy shock, εR,t = 0.01. According to (5),

log, terms of trade log, exports log, imports

Consider now a domestic economy risk premium shock, a jump in the

log, homogeneous output x 10

log, terms of trade log, exports log, imports

Next we consider a 0.01 innovation in log, government consumption, gt.

log, homogeneous output log, consumption log level, exchange rate

• Labor market in the previous model: