1.
IMPROVING THE EVIDENCE BASE ON
THE COSTS OF DISASTERS –
TOWARDS AN OECD FRAMEWORK FOR ACCOUNTING RISK
MANAGEMENT EXPENDITURES AND LOSSES OF DISASTERS
OECD High Level Risk Forum
Public Governance and Territorial Development Directorate
Catherine Gamper
November 21 2014, OECD Headquarters, Paris

2.
0
50
100
150
200
250
300
350
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012
AnnualeconomiclossesinUSDbillion
• Past decade: USD 1.5 trillion in economic damages from man-
made disasters (industrial accidents, terrorist attacks) and natural
disasters (primarily storms and floods)
• Increase in economic damages believed to outpace national DRR
investments…
• … though this claim cannot be supported by data as there is hardly
any available, especially on an internationally comparative level
• The development of standardised and comparable accounting
frameworks for DRM expenditure and disaster losses can:
– Support the evaluation of economic benefits of DRR investments
– Faciliate cross-ountry comparisons
– Systematic indicators on global DRR objectives could be built (to inform post
2015 SDG‘s, HFA2 etc.)
Why we need to better account for costs
of disasters
Source: EM-DAT: The OFDA/CRED International Disaster Database, Université catholique de Louvain, Brussels, Belgium, www.emdat.be
(accessed 14 November 2013).
Economic losses due to disasters in OECD
and BRIC countries, 1980-2012 (USD Billion)

3.
1. Review national and international efforts recording ex-post
disaster losses:
– Analyse their comparability
– Analyse strengths and weaknesses
→ provide basis for developing better methods, setting new international standard and
providing repository for such information
2. Assess ongoing national efforts and propose draft framework for
assessing public (and potentially private) spending for DRM:
– To better understand countries‘ expenditures ex-ante and ex-post of disasters
– Help policy makers understand whether their spending efforts lead to future
reductions in disaster losses
→ goal is to be sufficiently comprehensive so as to account for most such
expenditure items, while be broad enough to capture similar expenditures across
countries
Project Objectives

4.
1. Review of international and national data collection efforts and
methods on recording disaster losses
2. Review of international and national efforts to collect information
on public (and private) expenditure on disaster risk management
3. A draft framework to collect DRM expenditure information
including public and potentially private, across hazard and sectors
4. Going forward:
– test the framework in 2-4 OECD member countries
– refine framework
– mainstream it and establish OECD database in the long-term
Expected Outputs

5.
• Expenditure related costs somewhat easier to derive compared to
losses
• Losses are challenging to asses as they can be:
– Short, medium, or long-term
– Direct or indirect
– Not only felt locally, but trigger through economic sectors and countries
globally
• Identifying expenditure information across different departments
and sectors equally challenging:
– There are no central repositories for DRM expenditure information
– Multiple agencies and levels of government have DRM expenditures, each
their own way of describing this in budgets and national accounts
– Even more complex if expenditure for DRM is “embedded”
– Requires much effort and judgment to identify spending categories across
sectors and levels of government
Costs of expenditures vs. losses

6.
Australia
Bangladesh
Bolivia
Chile
Costa Rica
Egypt
Estonia
Ethiopia
Fiji Finland
France
Germany
Greece
Haiti
Honduras
India
Indonesia
Iran
Italy
Jamaica
Japan
KenyaMadagascar
Malawi
Mexico
Mozambique
Nepal
Netherlands
New Zealand
Norway
Pakistan
Philippines
Poland
Portugal
Slovenia
Thailand
Turkey
United Kingdom
United States
Venezuela
Yemen
0.5
1
1.5
2
2.5
3
3.5
2.7 3.2 3.7 4.2 4.7
AverageDeathTollperDisaster
1980-2013(log)
Real GDP per Capita, Year 2010 (log)
Significant decrease in fatality rates from disasters with increasing income
1980-2013
OECD Non-OECD
• Resilience against disasters in OECD countries is high , but higher
income countries still experience large economic losses
• Policy makers need a good understanding of past losses to face this
challenge and understand better whether their DRR investments
are effective
• A comprehensive account of how to measure economic losses can
be found in Meyer et al (2012); they distinguish:
– Direct tangible costs
– Losses due to business interruption
– Indirect costs
– Sometimes added: intangible costs
– To be fully comprehensive costs of reconstruction, recovery, planning and
implementation of risk prevention measures should be counted as well –
hardly included in international databases
Recording disaster losses
Source: Source: EM-DAT: The OFDA/CRED International Disaster Database, www.emdat.be - Université catholique de Louvain - Brussels - Belgium". Data for OECD and
BRIC countries (1980-2012). Figures are shown true to the year of the event. OECD Stat National Accounts GDP per capita in US$, constant prices, reference year 2005
Source: EM-DAT: The OFDA/CRED International Disaster Database, www.emdat.be - Université catholique de Louvain - Brussels - Belgium; OECD (2013),
“Gross domestic product (GDP) MetaData : GDP per capita, US$, constant prices, reference year 2005”, National Accounts OECD Statistics Database,
accessed on 14 November 2013, http://stats.oecd.org/

7.
Total losses…
… the sum of direct and indirect costs:
Source: Hallegatte and Przyluski, 2011.

8.
• None contain all cost categories, but efforts underway
to improve existing methods (notably EU, CRED):
• A number of international databases exist, but they
differ:
– In the hazards they include
– In the geographical coverage
– In the variables collected
– In the way economic losses are calculated
• Most common databases used are:
– EM-DAT
– DesInventar
– Swiss Re‘s SiGMA and Munich Re‘s NatcatSERVICE
Currently available INTERNATIONAL
data on disaster losses

9.
International loss data bases:
commonalities and differences
Year Event EM-DAT Des-
Inventar
NOAA
earthquake
database or
Storm Events
database
Dartmouth
Flood
Observatory
US
Sheldus
Swiss Re
explorer
Munich
ReError!
Reference
source not
found.
2005 Hurricane
Katrina
1 833 Does not
apply
24
Error! Reference
source not found.
1 053 875 1 836 1 322
2010 Chile
Earthquake
562 675 521 Does not apply Does not
apply
562 520
2010 Haiti
Earthquake
222 570 222 521 316 000 Does not apply Does not
apply
222 570 222 570
2011 Great East
Japan
Earthquake
19 846 Does not
apply
No estimate Does not apply Does not
apply
18 520 15 880
2012 Hurricane
Sandy
54 Does not
apply
68
Error! Reference
source not found.
65 No
estimate
237 210
Error!
Reference source
not found.
• Most include information on social (casualties, injured or affected
people) and physical/economic losses
• Although the sources they use are very similar (reports by national
governments, IO’s, or newspaper reports etc.), numbers can differ:
Number of Casualties recorded across different national and international ex-post loss databases

10.
Economic losses (in USD billion) recorded in:
International databases Other national or hazard-specific databases
Name of
database/Event
EM-DAT sigma NatCatService National Oceanic
and Atmospheric
Administration
(NOAA)
Dartmouth
Flood
Observatory
US Sheldus
Chile
Earthquake,
2010
30 33.28 (out of
which insured:
8.88)
30(out of which
Insured: 8)
30 Does not
apply
Does not apply
Hurricane
Sandy, 2012
50 73.78 (out of
which
insured:36.89)
65 (out of which
insured:30 )
Property damage:
24.91
No estimate No estimate
Hurricane
Katrina, 2005
125 173.44 (out of
which
insured:80.37)
125 (out of which
Insured: 62.2)
Property
damage:42.53;
crop damage: 1.93
60 Property
damage:74.27;
crop
damage:2.12
International loss data bases:
commonalities and differences
• Economic losses recorded in all major international loss databases
• Definitions of what is considered as “economic losses” is provided
by each of them, though it is not clear to what extent the recording
follows them because no primary data collection is undertaken
• Recorded loss amounts are very similar for some events and less so
for others:
→ It is difficult to
conclude that
differences are
due to different
economic loss
estimation
methods
COMPARISON OF ECONOMIC LOSSES
ACROSS INTERNATIONAL AND
SELECTED NATIONAL DATABASES

11.
National loss data bases
• Not consistently available across OECD members (OECD DAF survey for
details); sample countries used for this work:
– Japan, Australia, Canada, Italy, Slovenia, United States
• Variables included relatively similar (casualties, insurance costs, property
losses, damages to crops or livestock)
• Economic loss estimations only recorded by half
• Recordings are more consistent in national compared to international
databases (systematic questionnaires used etc.)
• Differences in governance:
– E.g. Slovenia organised centrally
– E.g. Italy different institutions for different hazard – no single multi-hazard loss database
• Some countries have exhaustive loss recordings for specific hazards, e.g.
Japan for floods
– Process similar to Slovenia
– Distinguishes public and private losses

12.
Ex-ante loss estimation methods
• Loss assessments can be carried out ex-post (actual loss assessment), but
also ex-ante
• Ex-ante estimations calculate potential impact of a certain type and
severity of disaster on potentially affected population and economy
• Can help assess avoided costs obtained through specific DRR investments
• Ex-ante assessments are usually conducted on a needs-basis and are not
systematically carried out and captured in databases
→ The OECD conducted an ex-ante
direct and indirect loss estimation of
a potential large-scale flood and its
ripple effects in the Paris
metropolitan area

13.
Assessing public (and private)
expenditure for DRM
Challenges to collect such information within and
across countries:
 Budgeting processes differ across and within countries depending on
political, fiscal or federal system
 Budgeting processes can be transparent and open and traceable for
anyone interested, or decided behind closed doors
 Little academic guidance available to carry out PE reviews
 Sectoral PE reviews relatively easy to obtain, but establishing this for the
same sector across countries more challenging

14.
Assessing public (and private)
expenditure for DRM
With regard to risk management some additional
challenges:
 In most countries no central unit responsible for co-ordinating risk
reduction activities
 Risk-related activities entail cross-sectoral expenditure items usually not
thematically reported in public accounts - may form part of public
spending in infrastructure, environment, planning sectors etc.
 Each sector may have its way of distinguishing hazards and types
investment (e.g. prevention, preparedness, rehabilitation).
 Expenditure may be “embedded”, i.e. expenditure for a project may only
partly pertain to risk reduction.

15.
Assessing public (and private)
expenditure for DRM
→ Challenge lies in first identifying disaster-related expenditure items in each
sector and categorise them broadly;
→ Based on this, comparable categories across sectors can be established to
aggregate specific expenditure categories across sectors, which can in turn
build the basis for a comparative framework across countries.
→ Retrieving private DRM spending also a challenge: information sits within
businesses and households – difficult to obtain systematically and over time
→ Outcome will be an approximation: establishing broad categories, based
on appropriate levels of aggregation and meaningful classifications of DRR
expenditures

16.
Assessing public (and private) expenditure for DRM –
what is the state of the art in OECD countries?
 Reviews that exist are result of specific project to retrieve DRM
expenditure information from national accounts and sectoral
budgets
 Usually one-off efforts (although some include historical data)
 Some focus on distinguishing expenditure along the DRM cycle (e.g.
preparedness vs. response spending), others gather information on
specific hazards
 A harmonised approach is needed – OECD could be useful vehicle
for inciting such data collection and building an international data
repository
 With this work OECD does not seek to change how public accounts
provide DRM expenditure information, the objective is to find a way
to obtain such information from governments in a comparative way
on a continuous basis

17.
Towards an OECD framework for
assessing public expenditure for risk
management

18.
Towards an OECD framework for assessing public
expenditure for risk management
 Based on the review of existing national efforts to obtain DRM
expenditure data it is suggested:
 To proceed along similar lines of developing a questionnaire and conduct
expert interviews in complement (as opposed to relying on public accounts for
this)
 To obtain expenditure information by :
 phase of DRM cycle so as to establish ex-ante versus ex-post DRM
expenditure,
 by level of government and hazard type

19.
Towards an OECD framework for assessing public
expenditure for risk management
1. Risk Prevention and Mitigation
(measures that decrease or eliminate
impacts on society and economy)
2. Preparedness
• Strategic Planning
• Hazard identification and assessment
• Risk/Hazard Mapping
• Land-use planning
• Planning, developing and constructing
protective infrastructure
• Prevention measures for the existing
built environment (houses etc)
• Prevention measures for critical
infrastructure (energy, water,
transport, road networks, ICT, etc.)
• Risk awareness and communication
activities
• Risk transfer investments by the public
sector
• Development of crisis management plans
• Early warning systems development,
construction and management
• Evacuation planning and management
• Emergency supply management
• Emergency preparedness/crisis management
exercises
The framework distinguishes:
 Expenditure items by phase of the DRM cycle (including dedicated
costs (directly related to DRM or embedded costs)):

20.
Towards an OECD framework for assessing public
expenditure for risk management
3. Emergency Response
(includes all expenses incurred in the
immediate aftermath of a disaster)
4. Rehabilitation and Reconstruction
• Emergency supplies
• Assistance packages to affected
regions, households etc.
• Payments to NGO‘s and other
emergency support agencies
• Expenditure related to immediate
response to public service disruption
(energy and water supply, transport,
etc.)
• Search and rescue operations
• Rehabilitation of public infrastructure
• Rehabilitation of private assets
The framework distinguishes:
 Expenditure items by phases of the DRM cycle:

21.
Towards an OECD framework for assessing public
expenditure for risk management
The framework distinguishes further:
 By type of actor that conducts investment:
 ministry/department
 national/sub-national or other (such as EU)
 private (households or businesses)
 By type hazard type
 Ideally by cost category:
 Staff costs
 Administrative costs
 Overheads
 Capital investment
 Operations & Maintenance
 Other

22.
Towards an OECD framework for assessing public
expenditure for risk management
Questions for discussion and feedback on the proposed framework:
 Is the framework capturing all relevant public (and private) expenditure
data?
 Could the distinction of phases be an agreeable distinction? Equally the
ex-ante and ex-post phases? This differs across countries and therefore it
is important to find agreement of what phases of the DRM cycle should be
distinguished and which phases fall under ex-ante or ex-post or whether
the latter is helpful at all?
 Is it feasible to distinguish cost categories (staff, administrative etc?)
 Is it feasible to distinguish expenditure items by risk type?
 Within the DRM phases are the sub-categories helpful? (we do not expect
countries to fill the sub-categories out, but rather see them as a check-list
of what should go into the aggregate category)

23.
For further information please contact:
catherine.gamper@oecd.org