Transcript Capital Flows, Interest Rates and Precautionary Behaviour

Capital Flows, Interest Rates and
Precautionary Behaviour: a model of
“global imbalances”
Marcus Miller and Lei Zhang
Discussion
Romain Ranciere (IMF)
usual disclaimer applies
• A very interesting paper that proposes an integrated framework to jointly
analyze two important questions:
– Global Imbalances: large US Current Account Deficit and low interest
rate
– Precautionary Motive for Reserves Accumulation
• a previous look on the link:
– view of reserves accumulation as “mercantilist” policies (Dooley and
al.)
but
– Aizenman-Lee (2006): more empirical support for precautionary
motive view than for mercantilist view: a more liberal capital account
increases international reserves.
• Models of precautionary reserves
– Caballero-Panageas (2005); Jeanne-Ranciere (2006)
– small open “emerging” economy
– no market for reserves assets; exogenous interest rate
– Insurance against “sudden stops”
– insurance domestic absorption/consumption against capital outflows [+
via balancesheet effect associated output losses)]
– Crisis prevention / Crisis smoothing
• Model of Global Imbalances and low-interest rates
– Deterministic Models
• e.g. Caballero and al. (growth differential; capital market
impefections)
– No risk; No demand for insurance
– Demand for insurance + Capital Market Imperfection  Self-insurance
Miller-Zhang Model(s)
• 2 “countries” (safe: US; risky: ROW); 2 period (risk in period 2)
• risk: mean preserving output risk. GDP insurance (vs. SS insurance)
• capital market (complete: insurance; incomplete: self-insurance)
• risk-aversion (standard log utility; loss-avoidance)
• intertemporal smoothing; crisis smoothing
• general result: the risky country saves i.e. is net buyer of insurance
securities (-> current account surplus in period 1). insurance premium ~
lower interest rate.
• remark: the 2-period case allows to mask an assumption on hard credit
constraint in bad times. (sudden stop literature: time-varying credit
constraint)
• Insurance
– complete market: contigent claims. sell goods in good state for goods in
bad states.
– remark: if the insurer was risk-neutral (and competitive) no
imbalance
– the insurer is risk-averse-> modest imbalance.[<0.2% of GDP]
• Self-insurance
– incomplete market: accumulates risk-free asset government bonds
– higher demand for US government bonds (lower interest rate)
– but (still) modest imbalance [<0.5% of GDP]
– why so little self-insurance?
– cost/benefit of insurance: intertemporal distorsion vs. between-state
distorsion. [increasing per unit cost of self-insurance]
• Risk-Aversion Issues
– log utility; constant relative risk-aversion~1.
– Jeanne-Ranciere (2006) dramatic shift in demand for insurance
between risk-aversion 1 and 4. [but more “crisis” insurance also means
less “intertemporal smoothing”]
– Natural extension: Epstein-Zin preferences separate preferences for
intertemporal substitution and risk-aversion.
• Loss Avoidance
– downside risk is more costly than upside gains.
– capture more broadly the “costs of crises”: political costs; distributional
issues  worth developing.
– alternative view: learning and overshooting misperception on the
economic costs of crisis
– RESULTS: combination of LA+severe output risk CAD 5% and
possibility of negative real interest rate
– Notice: here nominal = real one good. otherwise negative return 
US depreciation.
– US bonds is not a good insurance needs to buy a lot to get selfinsurance “war chess”  crises are specially costly  ready to pay
negative interest rate (cost of storage)
• comparison with Jeanne-Ranciere (2006): insurance against sudden stop
– Key difference: maturity transformation. the government accumulates reserves
because it is less credit constrained than private sector [it can borrow long term
while private sector borrows short-term]
– cost of a crisis is capital outflows (11%) and output cost (6.5%) 17.5%
reserves of 9.7%
– crises are rare and severe events (pi=10%); risk-aversion is higher (2). At
sigma=1 : zero reserves!
• Cost of Reserves: term premium. 1.5%
• Miller and Zhang: “excess” saving and insurance premium
remarks/questions
• somehow unresolved tension: sudden stops and global imbalances. sudden
stop typically occurs in the risky country with net capital inflows. who fears
a sudden stop now (US or emerging?)
• good simple model; possible extension towards calibration
– break-even more the symmetry (growth differential; different shocks,
different size paper by Imbs and Mauro (Pooling))
– real exchange rate depreciation
– estimation of the parameters (->time-varying parameter)
– OLG /simulation.
• global temporary phenomenon (temporary buildup...): yes and no:
probability of sudden stop increases with financial openness
• US big insurer? can the insurer fail? risk-less or just less-risky US
securities?