Transcript Timescale interactions in economy: Application to climate

Endogenous Business Cycles and the
Economic Response to
Natural Disasters
Stéphane Hallegatte
ENM, Toulouse and CIRED, Paris
With M. Ghil, P. Dumas, J.-C. Hourcade
Business cycles and extreme events
interactions
• Business cycles are interesting on their own
• For extreme events consequences, one have to
take into account preexisting economic
situation, including disequilibriums
Part I
Endogenous business cycles
The « real » cycle theory
• From Slutsky (1927) and Frisch (1933)
• First proposed by Kydland and Prescott (1982)
Business cycles and economic fluctuations arise
from exogenous shocks:
• Changes in productivity
• Changes in energy prices
• Fiscal shocks
• Aside from these exogenous shocks, the
economic system is stable
What does RBC theory assume ?
• Perfect markets : questionable, especially for
labor market over a few quarters
• No non-voluntary unemployment
• Rational expectations
• No role of quantities (e.g. aggregated demand)
(contrary to observations; Gali, 1999, 2004)
• Link with observable productivity shocks ?
Endogenous business cycles:
an alternative to real business cycles?
• In EBC models, cyclical behavior originates from
endogenous instabilities in the economic system
• It does not mean that exogenous shocks do not play
any role in economic fluctuations
NEDyM (Non-equilibrium Dynamic
Model)
• Represents an economy with one producer, one
consumer, one good to consume and invest.
• Developed from the Solow (1956) model, all
equilibrium constraints are replaced by dynamic
relationships with adjustment delays
• NEDyM equilibrium is neo-classical and identical
to the Solow model’s one.
• Because of market adjustment delays, NEDyM
dynamics exhibit keynesian features along
transients trajectory in response to shocks.
Distribution between physical
investments and dividends
• Managerial regime (Berle and Means, 1932):
the producer invests in all projects, up to a fixed level of
profitability.
• The remaining part of available liquidities are redistributed.
inv = 1.7
inv = 2.5
inv = 10
inv = 20
Hopf bifurcation:
from stable equilibrium, to limit cycle,
to chaotic behavior
Endogenous dynamics: an alternative explanation
for business cycles
Interpretation of the instability
• Business cycles originate from the profit-investment
relationship. Oscillations with a 5.5-year period.
higher profits > more investment > more demand > higher profits
• Business cycles constrained by 3 processes:
– Increase in labor costs when employment is high
– Constraints in production and inflation in goods
prices when demand increases too rapidly
– Financial constraints on investment
• A small additional instability arises from the interaction
of labor market and goods inventory (oscillations with a
300-day period)
Endogenous dynamics: an alternative
explanation for business cycles
• Harrod (1939): differences between population
growth and labor demand causes unstability
• Solow (1956): capital/labor substitution makes
economy stable
– even with capital/labor substitution (Cobb-Douglas
production function),
– even with neoclassical equilibrium,
– adjustment delays in markets (goods, labor, capital)
can cause endogenous economic fluctuations,
without exogenous shocks
• Good ability to reproduce some stylized facts: cycle
period, cycle asymmetry, variable correlations (e.g.
output and inflation)
Main flaw in the model business cycle:
the amplitude of price oscillations
Correction: accounting for quantities in
behavioral equations (Gali, 1999, 2004)
Part II
The economic consequences of
extreme events
Why to look at extreme events
in the context of climate change?
Natural disasters are an important part of the climate
change issue:
– Because of natural variability, climate change might
appear more through changes in the distribution of
extremes than through changes in mean conditions.
– Because mild changes in means can lead to large
changes in extremes and because the cost of
extremes is strongly non-linear with respect to their
intensity
Extreme events cannot be averaged
• Some economic models take natural disaster
costs into account through decreases in mean
productivity.

But natural disasters are rather a series of
shocks that destroy productive capital.
• This is equivalent only if the impact of disasters
can be averaged over long periods, i.e., if
impacts are linear, which is not the case.
How to look at extreme events?
• If disasters are modeled through capital destructions:



Using long-term growth model is impossible:
development of a disequilibrium model of economic
growth (NEDyM model)
Using a classical production function leads to
underestimated impacts on production (because of
decreasing returns)
Accounting for financial and technical constraints on
the reconstruction pace is necessary to reproduce the
observed response to past disasters over the short-
term.
Using a classical production function
would introduce a bias
Decreasing returns  when labor is fixed, capital is less and less
efficient, because the most efficient investments are realized first
If x% of capital is destroyed, production is reduced by less than x%
A modified production function
Investing less is not equivalent to having some productive capital
destroyed by a disaster
We introduce:
K0 = potential capital (no disaster)
K = portion of destroyed capital
Effective capital :
K = (1-K) K0
Modified production function:
Y = (1-K) ALK0(1-)
With this function, when x% of productive capital is destroyed,
production is reduced by x%
Modeling short-term constraints
Without taking into account short-term constraints: reconstruction is
carried out in a couple of month, even for large-scale events
At odds with observations (1999 winter storm over France, 2002
floods in central Europe, Katrina)
Technical constraints:
– availability of qualified workers, reconstruction organization, onsite accomodation of workers, material transportation and
availability
– Financial constraints (mostly developing countries):
We introduced a limitation of reconstruction investments at fmax
percent of total investments (calibrated on the length of the
reconstruction period):
Ir < fmax I
Long-term consequences on economic growth
Here also, the average loss depends on the event
characteristics and on reconstruction capacity
Long-term consequences on economic growth
Natural disasters and bifurcation in losses
From Hallegatte et al. (2007, Ecological Economics)
Index of disaster
frequency and intensity
Index of reconstruction capacity
The impact of a series of events is not the sum of the event impacts.
A macroeconomic feedback ?
Interaction between intrinsic dynamics and shocks
Recession
Expansion
From Hallegatte and Ghil (2007)
A macroeconomic feedback ?
Interaction between intrinsic dynamics and shocks
Recession
Expansion
Limited losses if the disaster
affects an economy in recession
From Hallegatte and Ghil (2007)
A macroeconomic feedback ?
Interaction between intrinsic dynamics and shocks
Recession
Expansion
Large losses if the disaster
affects an economy in
expansion
A vulnerability paradox: growing economies are more vulnerable !
From Hallegatte and Ghil (2007)
Variability and vulnerability
The cost of the current distribution of natural disasters strongly
depends on economic characteristics.
Long-term consequences on economic growth
« positive » feedback and poverty traps
Limited
reconstruction
capacity
« Positive » feedback
Long
reconstruction
period after each
disaster
Limited economic
development
Possibility of
poverty trap
Impossibility to
increase the
infrastructure and
capital stocks
Large cost of
natural disasters
Disasters can be a significant obstacle to economic growth and development