The Macroeconomic Consequences of Natural Disasters

1
The Macroeconomic Consequences of Natural
Disasters
Ilan Noy - University of Hawaii, Manoa
2
(No Transcript)
3
(No Transcript)
4
Economic Research on Disasters

• Applied micro (development) research
• consumption smoothing after disasters
• ex ante insurance
• risk-sharing
• Costs of individual events
• mostly to household incomes or labor and
  immigration choices
• less frequently also macroeconomic costs such as
  the fiscal cost of reconstruction
• Can we generalize from these results?

5
Macroeconomic Research on Disasters

• Albala-Bertrand (1993)
• 28 disasters
• Rasmussen (2004) the Caribbean
• Skidmore and Toya (EI, 2002)
• long-run impact of natural disasters
• cross-sectional dataset
• Yang (2006)
• ex post international financial flows
• more aid, more remittances
• Raddatz (JDE, 2007)
• explaining the importance of external shocks on
  output volatility in developing countries
• VARs including natural disasters

6
The Research QuestionsWhat are the
macro-economic consequences of natural
disasters?What determines these consequences?
7
In This Paper

• About 450 disasters
• Exact number depends on measure used.
• Panel data 1970-2003
• What is the short-run impact of natural
  disasters?
• What determines this impact?

8
EM-DAT Data

• 1900-2006
• 10 or more people reported killed
• 100 people reported affected
• declaration of a state of emergency
• call for international assistance
• Hydro-meteorological disasters
• floods, wave surges, storms, droughts, landslides
  and avalanches
• Geophysical disasters
• earthquakes, tsunamis and volcanic eruptions
• Biological disasters
• epidemics and insect infestations
• How accurate are these data?

9
EM-DAT Data

• For each disaster we know
• number of people killed (507)
• of population (t-1)
• number of people affected (466)
• of population (t-1)
• amount of direct damage (428)
• of GDP (t-1)
• onset month
• Disaster measure used

10
Disaster Variables
Variable Mean S.D. Min. Max. Max. Obs.
DDAMG 0.026 0.120 0.000 1.486 1.486 428
DKILP 0.085 0.654 0.000 11.047 11.047 507
DAFFP 5.438 12.686 0.000 98.767 98.767 466
Correlation DKILP- DAFFP 0.61 DKILP- DDAMG 0.32 DKILP- DDAMG 0.32 DKILP- DDAMG 0.32 DAFFP- DDAMG 0.50 DAFFP- DDAMG 0.50
11
The Specifications

• Basic equation
• Further hypotheses

12
Other Data on the RHS

• GDP growth (lag)
• GDP per capita (in PPP)
• CPI inflation
• Unemployment rate
• Population
• Institutional strength
• Investment growth
• Domestic credit
• Current account surplus
• Imports
• Foreign direct investment
• Financial crises

13
Hausman-Taylor Estimation

• Weighted instrumental variables estimation
• Like in the Arellano-Bond GMM, the instruments
  are deviations from country-specific means.
• Variances for coefficient estimates are obtained
  from a least-squares fixed-effects regression and
  a second-stage FGLS procedure.
• We specify endogenous/exogenous variables
• Classifying the disaster variables as either does
  not change the results.

14
Disaster Cost Regressions - Benchmarks
Crisis measure in monthly shares Crisis measure in monthly shares Crisis measure in monthly shares Binary crisis Binary crisis Binary crisis
Variable (1) (2) (3) (4) (5) (6)
DDAMG -86.956 -9.564
DAFFP 0.477 0.314
DKILP 743.336 0.314
GDPGL1 0.255 0.259 0.259 0.161 0.259 0.259
Observations 1574 1574 1574 1574 1574 1574
Adj-R2 0.13 0.13 0.13 0.13 0.13 0.13
F-test 3.02 2.96 2.96 72.93 2.96 2.96
15
Benchmarks by Size and Income Level
Variable (1) All (2) OECD (3) Developing (4) Small Econ (5) Big Econ
DDAMG -82.546 0.629 -84.894 -89.579 -44.231
GDPGL1 0.174 0.325 0.181 0.143 0.306
RR 0.043 0.047 0.073 0.072 0.041
GCFGL -0.002 -0.013 -0.003 -0.004 -0.002
DOMCREL -0.009 -0.010 0.001 0.005 -0.018
CAL 0.039 0.258 0.037 0.019 0.182
BUDGETL 0.000 -0.076 -0.003 0.040 -0.091
INFLL 0.000 0.004 0.000 0.000 0.000
IMPGL 0.011 0.026 0.008 0.008 0.009
FDIL 0.135 -0.038 0.138 0.150 0.092
SSHNG -2.339 -0.837 -1.600 -0.874 -1.627
16
Determinants of Disaster Costs Real variables
X (1) Illiteracy (2) Institutional strength (3) GDP per capita (4) Government consumption (5) Exports (6) Tropics
DDAMG 28.020 -48.246 -69.527 -149.446 -337.973 -162.37
DDAMGX -7.645 1.179 1.483 6.381 5.829 81.333
X 0.000 0.044 0.000 -0.078 0.031 -0.029
GDPGL1 0.105 0.177 0.164 0.160 0.180 0.182
RR 0.079 0.082 0.064 0.035 0.044
GCFGL 0.005 -0.003 -0.002 -0.002 -0.002 -0.002
DOMCREL -0.010 -0.009 -0.001 -0.010 -0.009 -0.009
CAL 0.062 0.039 0.053 0.036 0.036 0.040
BUDGETL -0.019 -0.002 -0.010 -0.013 -0.020 -0.003
INFLL 0.000 0.000 0.000 0.000 0.000 0.000
IMPGL 0.012 0.011 0.011 0.011 0.010 0.010
17
Determinants of Disaster Costs Financial
variables
(1) Stock market size (2) Domestic credit (3) Cap. acct. openness I (4) Cap. acct. openness II (5) For. exch. reserves
DDAMG -63.983 -216.378 -713.727 -471.740 -219.313
DDAMGX 147.943 228.482 -611.937 -98.289 71.252
X 1.002 -0.042 -0.109 -0.013 0.031
GDPGL1 0.130 0.164 0.176 0.198 0.176
RR 0.028 0.056 0.040 0.052 0.046
GCFGL 0.012 -0.002 -0.011 -0.006 -0.002
DOMCREL -0.022 -0.005 -0.009 -0.010
CAL 0.173 0.032 0.083 0.041 0.033
BUDGETL -0.079 0.011 -0.013 -0.007 -0.007
INFLL 0.000 0.000 0.000 0.000 0.000
18
The Consequences of Disasters

• Property damage does have an impact on ex post
  short-run output dynamics.
• The effects on the population do not seem to
  matter for post-shock macroeconomic dynamics.
• Developing countries and small economies are more
  vulnerable.

19
The Determinants of Costs I

• Countries are better able to withstand the
  post-disaster economic shock when they have
• higher literacy rate
• better institutions
• higher per-capita income
• larger public sector
• higher degree of openness to trade
• more foreign-exchange reserves
• higher levels of domestic credit
• Less-open capital accounts.

20
We did NOT find that

• the size of the agricultural or mining/minerals
  sectors matters.
• island economies are more vulnerable.
• large countries are less vulnerable.
• inflation increases following disasters
• because of monetized deficit spending or because
  of new credit creation coming from foreign-aid
  inflows.
• unemployment changes following disasters.
• results for trade and for investment were robust.

21
Caveats I - Exogeneity

• Ramcharan (JIE, 2007) and Raddatz (JDE, 2007) use
  disasters (and assume exogeneity).
• We use binary indicators.
• We use an alternative proxy for the magnitude of
  the disaster
• for earthquakes the Richter scale
• for storms wind speed.
• We examine the control variables whether there
  are any observable differences in their means
  between the disaster and non-disaster
  observations.

22
Caveats II

• Mismeasurement.
• Disasters often induce policy changes.
• Fiscal spending priorities and crowding out of
  capital expenditures.
• Lagged response of policy and therefore of
  disaster impact.
• Aid inflows may hide the true cost of a
  disaster.
• Different disaster types and risk probabilities
  will be associated with different degrees of ex
  ante preparedness and mitigation attempts.

23
Future Research

• Fiscal impact of disasters
• Neo-classical and endogenous growth
• What are the channels through which a disaster
  affects growth?
• Does insurance coverage matter?
• Present-value model of the current account
• Importance of capital account policies
• Best responders What are the right fiscal and
  monetary prescriptions?
• The role of foreign aid What is the impact of
  post-disaster aid surges?
• Climate change anyone?

24
(No Transcript)
25
Disasters (1968-2002)
Disaster type Number of disasters Killed per disaster Affected per disaster OFDA relief
Flood 1,675 170 1,724,851 0.22
Storm 1,175 646 601,490 0.17
Epidemic 737 249 27,528 0.12
Earthquake 559 1,522 173,015 0.21
Drought 326 18,657 5,740,623 0.30
Landslide 310 84 38,789 0.06
Fire 129 19 69,552 0.13
Cold wave 114 103 46,656 0.01
Volcano 102 853 39,008 0.27
Infestation 47 na 1,100 0.68
Food shortage 38 4,293 734,630 0.13
Eisensee Strömberg (QJE, 2007)
26
More Questions
27
Mitigation III
28
(No Transcript)