Implementing information on the costs and benefits of adaptation in a portfolio –based decision framework

Alistair Hunt Department of Economics,

University of Bath, UK

NCCARF, Australia June 29th 2010

Motivation for Research

• Essentially practical: Test possible tools for evaluation of adaptation

• Credible adaptation policy likely to depend on treatment of uncertainty as well as of economic efficiency (balance of costs and benefits), etc.

Uncertainties in climate impacts

Implications for Flood Management

• Riverine flood risks in Shrewsbury, UK– Adaptation

• Key problem: uncertainty in impacts may result in inappropriate level or type of adaptation

May be better to adopt a portfolio of options that reflect the decision-makers’ preferences relating to (economic?) optimisation versus reducing the chances of getting it wrong (variance from the “optimal”)

Use of currently used decision rule (CBA)

Implications for Flood Management

• Riverine flood risks in Shrewsbury, UK– Impacts

• Direct physical damage to residential and non-residential property

• Forgone output from short-term disruption to non-residential properties.

• Direct impacts on human health (mortality, injuries and mental stress).

– Uncertainties driven by CC and SE Scenarios as well as welfare cost estimation

Flood management decision-making: portfolio analysis

• Portfolio Analysis– utilises principle that since individual assets likely to

have different/unpredictable rates of return over time, an investor should ensure that she maximises expected rate of return and minimises variance + co-variance of asset portfolio as a whole rather than manage assets individually, (Markowitz (1952)).

As long as co-variance of assets is low then the overall portfolio risk in minimised, for a given rate of overall return.

Flood management decision-making: portfolio analysis

• economic efficiency criterion (Net Present Value) - principal determinant of portfolio return. Also measure NPV variance as indicator of uncertainty

instead of appraisal of single flood response options, a group of options are collectively appraised.

may be better able to capture variations in effectiveness of responses across a wider range of possible (climatic and socio-economic) futures.

NPV =

N

nnnn

N

nn

nN

nn

n

i

CB

i

C

i

B

000 111

Potential Flood Management Options Option Type Specific Options

Managing the Rural Landscape to reduce runoff

Rural infiltration

Rural catchment storage

Rural conveyance

Managing the Urban Landscape Urban storage

Urban infiltration

Urban conveyance

Managing Flood Events Pre-event measures

Forecasting and warning systems

Flood fighting actions

Collective damage avoidance

Individual damage avoidance e.g. property resistance

Managing Flood Losses Land use management

Flood-proofing

Land use planning

Building codes

Insurance, shared risk and compensation

Health and social measures

River Engineering River conveyance

Engineered flood storage

Flood water transfer

“Hard” defences

Economic welfare analysis of flood management options

• 3 options: hard defence; property resistance; warning system

• CBA for each option– Three degrees of implementation (20%, 50%, 100%)– Constant-scale economies in costs assumed– Four (consistent) CC/SE scenario combinations

• SE Scenarios only use popn + hhold size projections

– Portfolios created from combinations of two options and three options, each option disaggregated according to degree of implementation

Two-option Portfolio Analysis

0

2000

4000

6000

8000

10000

12000

14000

0 20000000 40000000 60000000 80000000 100000000

Variance

EN

PV

Three-option Portfolio Analysis

0

2000

4000

6000

8000

10000

12000

14000

0 10000000 20000000 30000000 40000000 50000000 60000000

Variance

EN

PV

Sample of results data for 2-option portfolios

Two-option portfolios

Hard Defence Warn System Total cost ENPV

Correlation Coeff Variance

% Imp Cost Share cost ENPV Cost Share cost ENPV

100 2,886 0.80 13,161 739

0.20 -463

3,625

10,384 0.99212764 92,813,436

50 1,443 0.65 13,553 790

0.35 -238

2,233

8,674 0.99118622 95,096,526

20 577 0.41 11,265 815

0.59 565

1,392

5,000 0.98708165 57,246,149

Results I

• Economic efficiency – variance trade-off (broadly) exists for both 2 and 3 option portfolios

• Sub-optimal portfolios can be identified

• Hard defences generally contribute most to higher NPV and higher variance; property resistance option has opposite effect.

Results II

• Pair-wise correlation coefficients are high:– costs of implementing individual options are

assumed not to change under different scenarios,

– benefits of each option are all modelled to increase according to the extent of climatic and socio-economic change

Conclusions I

• Adaptation assessment may be enriched by use of portfolio analysis – incorporates uncertainty more explicitly into decision-making.

• Reliant on reliable, quantitative data relating to both costs and benefits of identified adaptation options.

• Benefits of adopting portfolio analysis using economic efficiency criteria better demonstrated when options selected are more scenario-specific in design.

Conclusions II

• Future research priorities include:– Testing adaptation options that are more

scenario-sensitive e.g. in other contexts– Increasing numbers of options considered in

PA– Include dynamic dimension, incorporating

value of new information (quasi-option)– Applying portfolio analysis within a portfolio of

alternative decision rules