The Study of Economic Loss from Natural Disasters

Date: 30 Nov 2010
Speaker(s): Professor Leonard Smith,
Venue: London

The Munich Re Programme of the Centre for Climate Change Economics and Policy held an academic symposium, at which invited experts discussed findings and compared competing methodologies for normalising disaster losses.

The symposium provided an opportunity for presentations of latest findings and open, and progressive discussion on methodologies and the relevance of this research area to risk management planning more generally.

The symposium followed the publication of a new study by Eric Neumayer and Fabian Barthel on normalised disaster losses.

Aim of the symposium

The Munich Re Programme of the Centre for Climate Change Economics and Policy (CCCEP) held an academic symposium, at which invited experts discussed findings and compared competing methodologies for normalising disaster losses.

The symposium followed the publication of a new study by Eric Neumayer and Fabian Barthel on normalised disaster losses.

The symposium was held under Chatham House rules. A summary of the presentations and discussion follows.

Structure of the symposium

Session 1: State of the art of economic loss normalisation

Session 2: Future developments and the challenge of mitigation measures

Session 3: Beyond normalisation: the determinants of loss and policy implications

Concluding open roundtable discussion

Summary of presentations and discussions

Session 1: State of the art of economic loss normalisation

Chair: Leonard Smith (LSE)

Roger Pielke Jr. (University of Colorado at Boulder): ‘Disaster loss normalisation: methods, applications, limitations’

The literature on disaster loss normalisation has grown immensely over the past 15 years.

Sufficient experience exists to propose some general assertions about (a) robust methodologies, (b) applications in specific decision contexts, such as climate attribution and disaster risk management, and (c) the limitations of normalisation approaches, many of which are unlikely to be overcome.

Download PDF of Pielke’s presentation

Eberhard Faust (Munich Re): ‘On the robustness of conventional normalization using the example of US thunderstorm-related losses’

The talk’s title was not so much a statement than a perspective for investigating various parameter choices, using the classical approach of normalisation applied on US thunderstorm-related economic (and insured) losses.

This calculus arrives at a rescaling of past disaster losses according to the change in destroyable wealth in the area affected between the past year and today.

A comparison was made between the usage of various proxies for overall and insured wealth (GDP, housing stock, premia), different spatial resolutions and how we can learn something about drivers of trends in normalised losses from comparisons with time series of potential driving parameters.

Fabian Barthel (LSE): ‘Normalising loss from natural disasters: a global analysis’

The presentation presented new normalisation results for both insured and economic losses. Besides a global analysis, several region- and disaster-type-specific analyses were shown. Particular emphasis was on trends in disaster losses in Germany and the United States.

Download PDF of presentation

Session 2: Future developments and the challenge of mitigation measures

Chair: Robert Muir-Wood (Risk Management Solutions)

Ryan Crompton (Macquarie University): ‘When will anthropogenic climate change signals be detected in US tropical cyclone loss data?’

No anthropogenic climate change influence on Atlantic tropical cyclone behaviour and concomitant damage has yet been detected, though such an influence is projected in the future (Knutson et al. 2010).

Using projections from a recent study in the journal ‘Science’, it was shown that anthropogenic signals are very unlikely to emerge in a time series of normalised US tropical cyclone losses at time scales of less than a century.

This result extends more generally to global weather-related natural disaster losses.

Download PDF of Crompton’s presentation

Laurens Bouwer (Vrije Universiteit): ‘Historic and future extreme weather risks: what do we know currently and what could be improved?’

The risks of loss of life and damage to capital assets due to storms and floods are likely to become larger in the future.

The analyses of historic loss records indicate that increases from natural disasters observed over the past around the world are primarily caused by increasing exposure and value of capital at risk. Anthropogenic climate change has not had a significant impact on economic losses so far.

However, the records could potentially lead to a better understanding of the relationship between climate variations (i.e. natural variability) and losses, which may be further exploited.

Two case study examples in The Netherlands were used to show the impact of projected socioeconomic change on exposure, and the effect of climate change on disaster probabilities.

A comparison of various studies on projected future storm and flood risk shows that for the period up to 2040, the contribution from increasing exposure and value of capital at risk to changes in economic losses is likely to be much larger than the anticipated impacts from anthropogenic climate change.

More efforts are required to better incorporate the effects of changes in exposure, vulnerability and adaptation in impact studies. In particular, quantification of these causes is necessary to indicate their contribution to future change in risk.

Download PDF of Bouwer’s presentation

Patricia Grossi (Risk Management Solutions): ‘A modeller’s perspective on loss normalisation and mitigation’

In this presentation, the use of normalisation techniques in validating model estimates of loss was discussed, as well as the challenges in modelling mitigation measures.

Download PDF of Grossi’s presentation

Wolfgang Kron (Munich Re): ‘Potential influence of disaster reduction measures on losses’

Separating contributions to trends in natural catastrophe losses is a difficult task and the results involve high uncertainty. While most contributing factors tend to steepen the slope of an increasing trend, mitigation measures tend to attenuate it. In some cases (windstorm, earthquake) hardening of structures reduces their vulnerability while the hardening measures themselves are much more expensive to replace after having been damaged.

These measures, implemented over a long period (e.g. by a change in building code), follow a continuous, functional trend. In contrast to this, flood protection works are effective from a particular point in time, introducing a highly non-linear and even non-functional component; this is not only true for local and regional measures, but to some extent also on a nationwide and global scale. After completion of a flood defence structure, a trend may “go on the water-side of the dyke”, but it does not leave a loss footprint on the land side – at least as long as the design levels are not exceeded.

With a number of examples, the impact of various mitigation measures was shown, and their possible consequences for trends in losses revealed.

Session 3: Beyond normalisation: The determinants of loss and policy implications

Chair: Swenja Surminski (LSE)

Stéphane Hallegatte (CIRED – Météo-France): ‘Accounting for indirect impacts in loss estimation’

Most economic analyses of natural disasters focus on their direct economic impacts, i.e. on the value of what has been damaged or destroyed. Direct losses, however, represent only a fraction of total losses, and other mechanisms need to be taken into account, including output losses. This presentation will present concepts and methods for measuring and assessing indirect losses.

Eric Neumayer (LSE): ‘Economic loss versus human loss: what can be learned from the analysis of disaster fatalities?’

There are many studies which have tried to identify determinants of natural disaster fatalities, e.g. political regime type, corruption, level of economic development, disaster propensity, etc. Are these determinants likely to affect normalised economic loss from natural disasters as well?

Download PDF of presentation

Main points from open round-table discussion

There is an increasing number of studies which normalise economic losses. While older studies generally fail to detect a significant trend in normalised losses, a few newer studies (e.g. by Eberhard Faust and Neumayer & Barthel) find tentative evidence for increases in normalised losses.
There is a far-reaching consensus among participants that the new results cannot directly be attributed to climate change, given the limitations in data availability and quality.
The participants also agree about the importance of analysing in greater detail the underlying physical phenomena, i.e. normalised losses should mirror the frequency and intensity of natural disasters.
The participants disagree, however, about what can be learnt from the new results.
Some participants argue that the time series in the newer studies are simply too short to detect a signal of climatic change, as, for these periods of analysis, natural climate variability cannot be ruled out as the causing factor.
Other participants regard the new findings as first evidence for a potential effect of climate change on normalised natural disaster loss, but acknowledge that much more research is needed to corroborate the results.

Participants

Marloes Bakker: Netherlands Environmental Assessment Agency (PBL);
José Barredo: Land Management & Natural Hazards Unit, European Commission – Joint Research Centre;
Fabian Barthel: Department of Geography and the Environment, LSE;
Laurens M. Bouwer: Institute for Environmental Studies, Vrije Universiteit, The Netherlands
Ryan Crompton: Risk Frontiers – Natural Hazards Research Centre, Macquarie University, Australia;
Jan Eichner: Geo Risks Research/ Corporate Climate Centre (GEO/CCC), Munich Re;
Eberhard Faust: Geo Risks Research/ Corporate Climate Centre (GEO/CCC), Munich Re;
Patricia Grossi: Risk Management Solutions;
Stéphane Hallegatte: CIRED – Météo-France;
Philipp Hasenmueller: Geo Risks Research/ Corporate Climate Centre (GEO/CCC), Munich Re;
Gerd Henghuber: Group Communications – Strategic Topics, Munich Re;
Peter Höppe: Geo Risks Research/ Corporate Climate Centre (GEO/CCC), Munich Re;
Wolfgang Kron: Geo Risks Research/ Corporate Climate Centre (GEO/CCC), Munich Re;
Ana Lopez: Centre for Climate Change Economics & Policy, LSE;
John McAneney: Risk Frontiers – Natural Hazards Research Centre, Macquarie University, Australia;
Robert Muir-Wood: Risk Management Solutions;
Nicola Ranger: Centre for Climate Change Economics & Policy, LSE;
Eric Neumayer: Department of Geography and the Environment, LSE
Roger Pielke Jr: Center for Science and Technology Policy Research, University of Colorado at Boulder;
Ernst Rauch: Geo Risks Research/ Corporate Climate Centre (GEO/CCC), Munich Re;
Julio Serge: United Nations International Strategy for Disaster Reduction UNISDR;
Leonard Smith: Centre for the Analysis of Time Series, LSE;
Swenja Surminski: Centre for Climate Change Economics & Policy, LSE.