Transcript Discussion by F. Braggion

Comments on:
Gaytán and Ranciere
Banks, Liquidity and Economic
Growth
Fabio Braggion
Tilburg University and CentER
March 17, 2006
The Question
• How Banking System and Financial
Fragility evolves with GDP growth?
Rational Behavior of Economic Agents
determine the structure of the Banking
System
For certain levels of GDP agents may
prefer to have a fragile banking system
Model
• Based on Diamond and Dybvig (1981) and Cooper and
Ross (1998)
• Two periods overlapping generation model
• Young Work - Old make investment decisions and
consume
• Agents are subject to liquidity shocks. Early types and
late type. Types are private information
• Two technologies:
– liquid technology
– Illiquid technology (production)
• Like in Diamond and Dybvig… two equilibria
– Bank run and no-bank run
• Bank can design contract
– Allow bank run
– Give insurance: induce late types to tell the
truth
• What contract is best? It depends on the
stage of development
Preferred Result
For intermediate probability of a bank run
and high enough risk aversion:
– Low Income Countries prefer insurance
against bank runs
– Middle income countries prefer to exposed to
run
– High Income Countries prefer insurance
against bank runs
Intuition
• Trade off: take the gamble or be insured
• Agents are risk averse: they don’t like gambling
Max
(1  q )u cE   1   u cL   qu cR 
cR  w  k  hk 
• They try to get partially insured by accumulating more
liquid assets
Intuition
• Trade off: take the gamble or be insured
• Insurance has a cost: late types must tell the
truth. An IC constraint is binding and distorting
the economy
Max
u cE   1   u cL 
s.t.
cE  w  k  hk 
Intuition
• Low income. The economy is poor. The
accumulation of liquid assets imposes a big loss
on late types. Enforce the IC
• Middle income: The economy is richer, and the
loss for late types smaller. Expose the systems
to runs
• High income: (IC) distortion decreases with
income. Insurance gives a result close to first
best
Does it really apply?
• The result is sensitive to the parameters
values
• What does it mean to have intermediate
probability?
• And also curvature of the utility function
Possibility to get a number:
Sapienza, Guiso and Zingales (2005)
Check out
• Also young agents are risk averse
• Decisions of old agents determine their endowment
at the beginning of their second period
(1   ) ( wt 1 )  1   wt 1  f k ( wt 1 )  with

wt  
(1   )f k ( wt 1 )  with prob (1   )

prob

• For high enough risk aversion there is the possibility
of an intergenerational transfer (young promising a
payment to the old) that induces the old to choose
insurance
Towards a Theory of Total Factor
Productivity
• By accumulating liquidity banks make the existing stock
of capital more efficient
 ( wt 1 )  1   wt 1  f k ( wt 1 )  with

Yt  
f k ( wt 1 )  with prob (1   )

prob

• Banks avoid inefficient liquidation of physical capital
• On the empirical side:
Benhabib and Spiegel, 2000
Use the set up of this model
• To explain TFP declines associated to
Financial Crises
(Meza and Quintin, 2005)
• Notable application: The great depression
In the spirit of Bernanke, 1983
Growing through phases
• Both with covered and uncovered banking
system growth is driven by:
– Capital accumulation for low levels of wealth
– Total factor productivity growth for
intermediate-low level of wealth
– Capital accumulation (again) for higher level
of wealth
Parallel with Matsuyama “Growing through
Cycles” 1999
But different channel:
while wealth increases, the financial
system becomes more efficient and copes
better with liquidity shocks
A dynamic model of institutions?
The optimal contract evolves with the state variable
Novel features
If you call the contract “institution”, we may have a
dynamic model of institutions