IMF Pillar Summary 20/21
IMF Pillars
Pillar 1: Exchange rate determination models 1
The Dornbush model 1
The Flex Price model 6
The Mundell-Fleming model 11
Pillar 2: Theories around the exchange rate 13
PPP (Purchasing Power Parity) 13
CIP (Covered Interest Parity) 15
UIP (Uncovered Interest Parity) 17
RP (Risk Premium) 19
Swam Diagram (1995) 19
Elasticity approach (to the BoP) 22
BoP (Balance of Payments) 23
Forex 29
Pillar 3: Stories/ IRL Situations 31
Bretton Woods 31
History of the Eurozone 31
OCA 32
Hamada / Currency policy coordination 32
Fixed vs. Flexible Exchange rates 32
Latin-American debt crisis 36
Pillar 1: Exchange rate determination models
The Dornbush model
Motivation from empirical observations (can the model explain these?):
1. Goods prices are sticky, e.g., because wages are sticky and resistant to
downward pressure: yes (provided π<∞)
2. Prolonged departures from PPP: yes, without irrational behavior
3. Exchange rates are more volatile than goods prices: yes
4. Exchange rate can overreact after a shock (overshoot new equilibrium): yes
5. Capital flows are important for exchange rate: yes, disequilibrium in
6. the bond market drives the exchange rate.
7. Interest rate clears the money market: yes
General information:
● Idea: (Consider a change in an exogenous variable (such as m(supply),y)
○ Money market equilibrium is disturbed: r clears
○ Δr disturbs int. bond market: capital flows affect s and Es• till
equilibrium
1
,IMF Pillar Summary 20/21
○ Δs disturbs goods market: p changes a bit
○ ∆p disturbs money market: r clears
○ Δr disturbs ..... until the long-run is achieved.
Assumptions
● Money market is cleared by r
○ Demand:md = p + ηy – σr
○ Supply:ms = m
○ Equilibrium:md = ms (permanent)
■ m = p + ηy – σr.
○ Comparison to flex price model: these assumptions are the same but
the market-clearing variable is r instead of p.
● International bond market is cleared by s
○ Equilibrium (UIP): Es• = r-r* (permanent)
○ Where expectations are regressive (cheap currency is expected to
appreciate):
■ Es•=(s---s), where s--=S is the long-run exchange rate.
■ Mind the bar!
○ Comparison to the flexprice model: these assumptions are the same
except for:
■ Market-clearing variable: s instead of r
■ Expectation Es• depends on s instead of Ep•-Ep*•
● Domestic goods market is cleared by p
○ Phillips curve: :p•= π(d-y) (π is degree of price flexibility)
○ Demand: d=β+α(s-p+p*)+φy
■ (demand for home-produced goods; drop λr from Pilbeam
(7.15): simpler. s-p+p* is the real exchange rate. We assume
φ<1.
○ If you prefer using d=s-p+p*+0.5y, that is also fine.
○ Only in the long run equilibrium : d=y
○ Comparison to the flex price model: these assumptions are new,
because it is not necessary to specify what drives p in the short run.
● PPP is not an assumption:
○ In long-run: d=y real exchange rate s---p--+p* = [(1-φ)y-β]/α.
○ Implications for real exchange rate (RER)
■ RER is nonzero in general
■ RER can change after real shocks.
● e.g., higher productivity causes an increase in y↑
● thus as φ<1: s---p--+p*↑ so that the foreign country
will consume extra domestic output.
○ Long-run PPP (s--=p---p*) is no assumption of this model.
○ Note:
■ Monetary shock does not change s---p--+p*; still, that is not
an assumption, but a result.
2
,IMF Pillar Summary 20/21
■ Productivity shock changes s---p--+p*, so long run of
Dornbusch is not the same as flexprice model in every respect.
● Dornbusch as a “special case”of UIP
○ UIP can be the basis for an exchange rate theory, i.e., st is determined
by rt & Et{st+1} (taking r*t as given).
○ The Dornbush model can be viewed as a specification of both:
■ r comes from money market equilibrium
■ Es comes from Ө(s---s)+s,
■ where s-- comes from long-run goods market equilibrium
○ Notice how all 3 market assumptions & regressive exp. formulas are
needed.
○
Graph:
● 3 Endogenous variables (p,r,s)drop one (r) in 2-dim. graph: s,p-graph.
● 3 Market equilibria to be represented in lines:
○ Domestic goods market: depends on p&s, not r fine in s,p-graph: GG
line
○ Money market:: depends on p&r, not s difficult in s,p-graph: Δr shifts
line
○ International bonds market: Internat. bond : depends on r&s, not p
○ difficult in s,p-graph: Δr shifts line
○ We combine the money and international bond market by
substituting out r
■ This gives us the money and bond markets that depends on p&s
this is fine in the s,p-graph this is the MB line (MM)
● We thus have a graph with two equilibrium lines.
● Slope of the GG (goods market equilibrium) line:
○ Starting from equilibrium (p,s): s increases and causes net exports to
increase which causes d to increase and thus d>y. To restore goods
market equilibrium (i.e., decrease net exports), p must increase): GG
is upward sloping.
3
, IMF Pillar Summary 20/21
○
● Slope of the MB line:
○ Starting from equilibrium (p,s):
○ p↓ thus md↓: md <ms so disequilibrium in money market
■ to restore equil., r must ↓ (no need for change in s yet)
■ disequilibrium in internat. bond market: excess demand for F
bonds.
○ To restore equilibrium Esdot must decreases and thus s must
increase: MD is downward sloping
○
● Entire graph:
●
○ Do not need to remember the formulas on the graph.
How does a shock work?
● Exchange rate overshooting after an increase in money supply (m):
○ The money supply shock causes MB to move upwards ( as the increase
in p restores the money market equilibrium and does not disturb UIP)
○
4
IMF Pillars
Pillar 1: Exchange rate determination models 1
The Dornbush model 1
The Flex Price model 6
The Mundell-Fleming model 11
Pillar 2: Theories around the exchange rate 13
PPP (Purchasing Power Parity) 13
CIP (Covered Interest Parity) 15
UIP (Uncovered Interest Parity) 17
RP (Risk Premium) 19
Swam Diagram (1995) 19
Elasticity approach (to the BoP) 22
BoP (Balance of Payments) 23
Forex 29
Pillar 3: Stories/ IRL Situations 31
Bretton Woods 31
History of the Eurozone 31
OCA 32
Hamada / Currency policy coordination 32
Fixed vs. Flexible Exchange rates 32
Latin-American debt crisis 36
Pillar 1: Exchange rate determination models
The Dornbush model
Motivation from empirical observations (can the model explain these?):
1. Goods prices are sticky, e.g., because wages are sticky and resistant to
downward pressure: yes (provided π<∞)
2. Prolonged departures from PPP: yes, without irrational behavior
3. Exchange rates are more volatile than goods prices: yes
4. Exchange rate can overreact after a shock (overshoot new equilibrium): yes
5. Capital flows are important for exchange rate: yes, disequilibrium in
6. the bond market drives the exchange rate.
7. Interest rate clears the money market: yes
General information:
● Idea: (Consider a change in an exogenous variable (such as m(supply),y)
○ Money market equilibrium is disturbed: r clears
○ Δr disturbs int. bond market: capital flows affect s and Es• till
equilibrium
1
,IMF Pillar Summary 20/21
○ Δs disturbs goods market: p changes a bit
○ ∆p disturbs money market: r clears
○ Δr disturbs ..... until the long-run is achieved.
Assumptions
● Money market is cleared by r
○ Demand:md = p + ηy – σr
○ Supply:ms = m
○ Equilibrium:md = ms (permanent)
■ m = p + ηy – σr.
○ Comparison to flex price model: these assumptions are the same but
the market-clearing variable is r instead of p.
● International bond market is cleared by s
○ Equilibrium (UIP): Es• = r-r* (permanent)
○ Where expectations are regressive (cheap currency is expected to
appreciate):
■ Es•=(s---s), where s--=S is the long-run exchange rate.
■ Mind the bar!
○ Comparison to the flexprice model: these assumptions are the same
except for:
■ Market-clearing variable: s instead of r
■ Expectation Es• depends on s instead of Ep•-Ep*•
● Domestic goods market is cleared by p
○ Phillips curve: :p•= π(d-y) (π is degree of price flexibility)
○ Demand: d=β+α(s-p+p*)+φy
■ (demand for home-produced goods; drop λr from Pilbeam
(7.15): simpler. s-p+p* is the real exchange rate. We assume
φ<1.
○ If you prefer using d=s-p+p*+0.5y, that is also fine.
○ Only in the long run equilibrium : d=y
○ Comparison to the flex price model: these assumptions are new,
because it is not necessary to specify what drives p in the short run.
● PPP is not an assumption:
○ In long-run: d=y real exchange rate s---p--+p* = [(1-φ)y-β]/α.
○ Implications for real exchange rate (RER)
■ RER is nonzero in general
■ RER can change after real shocks.
● e.g., higher productivity causes an increase in y↑
● thus as φ<1: s---p--+p*↑ so that the foreign country
will consume extra domestic output.
○ Long-run PPP (s--=p---p*) is no assumption of this model.
○ Note:
■ Monetary shock does not change s---p--+p*; still, that is not
an assumption, but a result.
2
,IMF Pillar Summary 20/21
■ Productivity shock changes s---p--+p*, so long run of
Dornbusch is not the same as flexprice model in every respect.
● Dornbusch as a “special case”of UIP
○ UIP can be the basis for an exchange rate theory, i.e., st is determined
by rt & Et{st+1} (taking r*t as given).
○ The Dornbush model can be viewed as a specification of both:
■ r comes from money market equilibrium
■ Es comes from Ө(s---s)+s,
■ where s-- comes from long-run goods market equilibrium
○ Notice how all 3 market assumptions & regressive exp. formulas are
needed.
○
Graph:
● 3 Endogenous variables (p,r,s)drop one (r) in 2-dim. graph: s,p-graph.
● 3 Market equilibria to be represented in lines:
○ Domestic goods market: depends on p&s, not r fine in s,p-graph: GG
line
○ Money market:: depends on p&r, not s difficult in s,p-graph: Δr shifts
line
○ International bonds market: Internat. bond : depends on r&s, not p
○ difficult in s,p-graph: Δr shifts line
○ We combine the money and international bond market by
substituting out r
■ This gives us the money and bond markets that depends on p&s
this is fine in the s,p-graph this is the MB line (MM)
● We thus have a graph with two equilibrium lines.
● Slope of the GG (goods market equilibrium) line:
○ Starting from equilibrium (p,s): s increases and causes net exports to
increase which causes d to increase and thus d>y. To restore goods
market equilibrium (i.e., decrease net exports), p must increase): GG
is upward sloping.
3
, IMF Pillar Summary 20/21
○
● Slope of the MB line:
○ Starting from equilibrium (p,s):
○ p↓ thus md↓: md <ms so disequilibrium in money market
■ to restore equil., r must ↓ (no need for change in s yet)
■ disequilibrium in internat. bond market: excess demand for F
bonds.
○ To restore equilibrium Esdot must decreases and thus s must
increase: MD is downward sloping
○
● Entire graph:
●
○ Do not need to remember the formulas on the graph.
How does a shock work?
● Exchange rate overshooting after an increase in money supply (m):
○ The money supply shock causes MB to move upwards ( as the increase
in p restores the money market equilibrium and does not disturb UIP)
○
4
