Download THe NK Approach to Exchange Rate Policy Analysis: Looking Forward

THe NK Approach to Exchange Rate Policy
Analysis: Looking Forward
Tommaso Monacelli - Università Bocconi, IGIER and CEPR
The Transmission of International Shocks to Open Economies
Reserve Bank of New Zealand, December 2010
NK analysis of optimal monetary policy in open economies
(OE)
I
Fundamentally di¤erent from closed economy counterpart
I
Baseline SOE model has two features
1. complete exchange rate pass-through
2. frictionless (domestic/international) …nancial markets
Optimal Monetary Policy in NK Models
I
Closed economy: in‡ation (markup) stabilization
I
Open economy: domestic markup stabilization is not e¢ cient
1. Via variations in international relative prices (terms of trade
and/or real exchange rate), can improve upon the
‡exible-price allocation
2. General nature of openness: variations in relative prices can
a¤ect consumption for any given level of output (and
therefore labor e¤ort).
Divine coincidence in the baseline model
I
Under log consumption utility, need special value of trade
elasticity (=1)
I
Income and substitution e¤ects of terms of trade movements
cancel out
Literature
I
Clarida Galí and Gertler (2001), Devereux and Engel (2003),
Benigno and Benigno (2003, 2006), Galí and Monacelli
(2005), Corsetti and Pesenti (2001, 2005), Faia and Monacelli
(2008), Corsetti et al. (2009), De Paoli (2009)
I
Recently
1. Engel (AER, 2010)
2. Corsetti, Dedola and Leduc (2010, Handbook chp.)
Openness and optimal PPI in‡ation volatility
Benchmark model (Faia and Monacelli 2008)
Two deviations from the baseline model
1. Local Currency Pricing
2. Imperfect Financial Markets
Local Currency Pricing
Breaking the divine coincidence in OE: LCP stability of
import prices
I
Deviations from the law of one price
1. At consumer level (Engel 010, CDL 010) (consumer-LCP)
I
Sources: (i) price stickiness ; (ii) local real factors
(distribution costs)
2. At the dock
Two main implications of "consumer LCP"
1. Should care about "currency misalignments" (Engel 2010) /
law of one price gap (Monacelli, 2006)
pF ,t
ψt
et
et + pF ,t
log Et
!Deviations from the law of one price as endogenous cost-push
shocks.
π H ,t = βEt fπ H ,t +1 g + κ y yet +κ ψ ψF ,t
|{z}
|{z}
output
gap
LOP
gap
2. Should care about CPI in‡ation
Quasi-divine coincidence? (CDL, 010)
But "dock-LCP" even more pervasive
1. Import prices very sticky at the dock (Gopinath and
Rigobon, 2007)
2. Stickiness of import prices higher for more di¤erentiated
goods (G-R, 2007)
3. Import price rigidity has increased by 10 percentage points in
1994-2005 (G-R, 2007)
4. U.S. import prices with high frequency of price adjustment
have a higher long-run pass-through (Gopinath and
Itskhoki, 2009)
5. Pass-through of the average good priced in dollars is 25% vs.
95% for non-dollar priced
Downward Trend in the Frequency of Price Adjustment
I
Decomposition (G-R 07)
∆P stickiness = ∆(N di¤er. goods) + ∆(P.stickiness di¤er. goods)
It is NOT a Compositional Story
1. Main suspect: increased degree of stickiness in prices of
di¤erentiated goods
2. Need a new story linking: " trade $ "price stickiness in
di¤erentiated goods
Some facts about exchange rates and asset prices in SOEs
1. Equity prices and real housing prices strongly co-move
2. Real ex. rates and housing prices (very) positively correlated
3. Real ex. rates and equity prices positively correlated, but less
obvious across countries
Real Exchange Rate Decomposition
RER t
=
Et PT ,t
PT ,t
= RER T ,t
(PT ,t /PN ,t )1
PT ,t /PN ,t
ω
1 ω
RER N ,t
I
Mechanical e¤ect of housing price variations on RER N ,t ?
I
Role of credit (booms)
Imperfect Financial Markets
Recent contributions
I
Benigno (2009), De Paoli (2010), Corsetti, Dedola and Leduc
(Handbook chp. 2010), Devereux and Sutherland (2009)
Financial imperfections in these papers
!Financial autarky
!Bond economy
I
Insight: "consumption imbalances" (CDL) generate wedge in
the risk-sharing condition
I
Under some conditions, can exist a tradeo¤ between (i)
domestic markup stabilization and (ii) "consumptionwedge" stabilization
Quasi-divine coincidence again?
I
But in all cases, frictions assumed exogenously
Looking forward
1. Making …nancial imperfections endogenous
2. An integrated approach to optimal policy analysis
Two types of frictions
1. Limited commitment
2. Adverse selection
I
Both types of frictions can become endogenously binding
Limited commitment
I
Typically at the core of models such as Kiyotaki and Moore
(1997), Cooley, Marimon and Quadrini, Quadrini and Jermann
(2009).
I
But also Bernanke and Gertler (1989), Bernanke, Gertler and
Gilchrist (1999), Carlstrom and Fuerst
I
More recently: Christiano et al. (2008), Gertler and Kiyotaki
(2010).
Example for a SOE
I
Households
(
Et
Xt =
Ct =
∞
∑ βt U (X,t , Nt )
t =0
1
θ
θ 1
θ
1
η
η 1
η
γ Ct
+ (1
α Ch,t + (1
)
1
θ
θ 1
θ
1
η
η 1
η
θ
θ 1
γ) Dt
α) Cf ,t
η
η 1
I
Sequence of budget constraints:
PC ,t Ct + PD ,t (Dt
I
Dt
1 (1
δ)) + Et Rt
1 Bt 1
= Wt Nt + Et Bt + τ t
Limited commitment
!Cannot commit today to repay tomorrow more than the
expected value of collateral
Rt Bt
| {z }
in
foreign
currency
(1
χ ) Et
PD ,t +1 Dt
E t +1
A Simple Theory of the Credit Spread
I
Pseudo-Euler condition
λt = βEt
I
λt +1 Rt
E t +1
Et
E¤ective real interest rate
mt
λe,t
βe λe,t +1
+
ψt
λt Rt
Et
Credit spread st
st
E t +1
Et
Et +1 ψt (Rt /Et )
= Rt
Et 1 ψt (Rt /Et )
ψt
E t +1
Rt
Et Et /Rt ψt
=
Et
mt
Rt
I
Increasing in the shadow value of borrowing ψt
I
Increasing in the expected exchange rate depreciation
Endogeneity of frictions
1. Value of collateral depends on asset price / ex. rate
‡uctuations
2. Borrowing constraint only occasionally binding
Extensions (1)
I
Can add a portfolio problem: optimal share of borrowing in
domestic/foreign currency
Rt Bt
Rt Bt
I
(1
(1
χ)Et fPD ,t +1 Dt g
χ )Et
PD ,t +1 Dt
E t +1
Link to growing literature on portfolio choice in general
equilibrium
Devereux and Sutherland (2008), Tille and Van Wincoop (2008),
Courdacier and Kollmann (2009).
I
Role of durables in trade (Engel and Wang 2010)
Integrated approach to optimal policy
I
Current approach: optimal stabilization policy conditional on
…nancial imperfections being in place
!Idea: there is an optimal policy for normal times vs. optimal
policy for turbulent times
I
But optimal policy in normal times should internalize that
…nancial imperfections may become binding in the future
I
Avoid bening neglect of asset price booms + credit booms
+ bubbles
I
Business cycle literature of the crisis should learn more from
the sudden stops literature
I
References: Mendoza (2010), Mendoza and Smith (2006)
Adverse Selection
I
Especially relevant during the crisis
I
Focus: debt and liquidity in interbank markets
Debt overhang
Figure: Total Credit Market Debt Owed: US Sectors
Debt markets subject to spirals
Liquidity and cost of …nance
I
Spirals involve feedback e¤ects between the cost of …nance
and liquidity in the interbank market
I
Liquidity = perceived safety of the assets traded in the
market.
I
Holmstrom (2008): necessary condition for asset to be liquid
is that market participants know its value.
Ampli…cation
I
Bad news hit ! informational asymmetries become more
acute ! Market participants cast doubts on the value of the
assets
I
Two e¤ects
1. Asset prices fall
2. Asset liquidity shrinks
=)Debt markets become "markets for lemons" (Tirole 2010)
=)It’s adverse selection
Banks
I
Some important elements
1. "Shadow banks" and interbank/repo market
2. Maturity mismatch and liquidity problem
3. The "right" …nancial shock
Financial intermediaries
! Current crisis: key role of balance-sheet e¤ects of banks
Key element in the crisis
I
Liquidity problem for "new" …nancial intermediaries
traditional banks
"investment" banks
Assets
long-term loans
MBS
Liabilities
deposits
short-term debt
Boom of securitized products
I
Used as collateral in interbank/repo markets
Fraction of AAA rated securities
securitized products
60%
corporate bonds
1%
(source Fitch, 2007)
Gigantic maturity mismatch
I
Banks held long-term assets (e.g., MBS) …nanced via
short-term debt (e.g., commercial paper)
I
When things deteriorate it is the liquidity problem that
matters
The role of securitization
I
Diversi…es idiosyncratic risk, but increases sensitivity to
aggregate risk
The magic of securitization
Coval, Jurek, and Sta¤ord (2010)
I
I
Suppose two identical bonds, each with probability of NOT
default = 0.9 ! prob. default = 1 0.9 = 0.1
NB: prob. default uncorrelated
I
Combine them in a CDO (collateralized debt obligation)
1. Junior tranche: pay if both tranches do not default
2. Senior tranche: defaults only if both default
junior
PAY
0.92 = 0.81
DEFAULT
1 0.81 = 0.19
senior
0.99
(1
0.9)2 = 0.01
I
Result: credit enhancement for the senior tranche
I
"Side e¤ect": tranches become correlated even if underlying
assets are not
Dynamics of crisis
Bad shock (what is this? )
=)Financial conditions deteriorate
=)Lenders reduce exposure !Ask to service debt
=)Banks try to …re sale long-term illiquid assets
Liquidity friction
I
At least as crucial as borrowing friction
I
Is it "…re-sale per se" or is it "…re sale" of long-term illiquid
assets"?
! Requires modelling of:
(i) Interbank market
(ii) Maturity of assets
Sketch of a model with two ingredients
I
"Banks sudden stop"
I
Shock to haircut margins
Agents
I
Savers
cs ,t +
dt = rtd 1 dt
|{z}
1
+ wt
deposits
I
Entrepreneurs
ce,t + qt (ke,t
ke,t
1) +
rtl (m )lt (m )
|
1 m l
rt
m k∑
=1
(1
k (m )lt k (m )
χ)Et fke,t +m
{z
collateral constraint
=
lt (m )
| {z }
long-term loans
1 qt +m g
}
+yt
Banks
I
Commercial banks
bt + Φ(bt ) = dt
Banks (con’t)
I
Investment banks
V (bt
1 , lt k , γt )
= max [xt + βx Et fV (bt , lt , γt +1 )g]
xt = bt +
1 m l
rt
m k∑
=1
xt ,b t ,lt
k (m )lt k (m )
lt (m )
s.t. haircut constraint:
rtb bt
|{z}
short-term borrowing
I
(1 γt ) lt (m)
| {z }
haircut shock
Banks’sudden stop: haircut constraint becomes
(endogenously) binding