investment risk modelling

48
Clean water for Wajir – building a business case for a pipeline Eike Luedeling

Upload: world-agroforestry-centre

Post on 24-Jan-2015

161 views

Category:

Economy & Finance


0 download

DESCRIPTION

World Agroforestry Center Production

TRANSCRIPT

Page 1: Investment Risk Modelling

Clean water for Wajir – building a business case for a pipeline

Eike Luedeling

Page 2: Investment Risk Modelling

Decisions under uncertainty

All decisions are made under uncertainty

We rarely know enough to be completely sure about how to decide

Decisions must be made, so we need strategies for making the best possible decisions under uncertainty

Image source: http://www.mynamesnotmommy.com

Page 3: Investment Risk Modelling

Uncertainty around the pipeline

Will the project succeed?

Will better water spur growth in Wajir?

Will Habaswein’s water supply be affected?

How much will water cost and how many people will buy it?

…Image source:

http://www.accountancyage.com

Page 4: Investment Risk Modelling

Decision modeling

Construct impact models for a decision

Identify all uncertain variables

Run simulations for many plausible scenarios that could happen

Summarize results for an idea of the range of likely outcomes

Image source: http://ronzopleaj.wordpress.com/

Page 5: Investment Risk Modelling

Decision modeling

Initial stakeholder workshop in November 2013 for joint framing of the decision.

Page 6: Investment Risk Modelling

Decision modeling

Two days of model development with core modeling team.

Page 7: Investment Risk Modelling

Decision modeling

Coding of the model

Incorporation of feedback

Elicitation and consolidation of estimates

Simulation runs

Identification of important variables

Image source: http://www.dignited.com

Page 8: Investment Risk Modelling

The model

Residents of Wajir

Residents of Habaswein

Residents along the pipeline

Upstream areas

Downstream areas

Water company

Donor

Separate calculations for all

stakeholder groups

Summed up to produce overall

project outcomes

Stakeholders

Page 9: Investment Risk Modelling

The model

Reduced infant mortality

Reduced costs for disease treatment

Higher productivity

Job creation

Higher investments

Reduced brain drain

Revenue from more demand for local products and services

Higher tax revenue

Reduced reliance on shallow wells

Sanitation benefits

Water during the dry season

General livelihood improvement

PES

Revenue from water sales

Benefits

Page 10: Investment Risk Modelling

The model

Initial investment

Running costs /metering

Salaries

Repairs

Aquifer monitoring

Infrastructure maintenance

Pipeline security

Payments for ecosystem services

Environmental impact

Costs for water purchases

Costs

Page 11: Investment Risk Modelling

The model

Negative feasibility report

Water yield too low

Inadequate benefit sharing

Political interference

Wells run dry

Increasing water

salinity

Oil development (raising the risks of wells running dry or turning saline)

Dam development (raising the risks of wells running dry or turning saline)

Maintenance

problems

Water price is too high

Regional conflict

Poor maintenance and operation

Illegal abstractions

Lack of cooperation

Risks

Page 12: Investment Risk Modelling

The model

Costs and

benefits

Success or

failure

Losses through inefficie

ncies

Time preference adjustmen

t (discounti

ng)

Outcome

Model structure

Page 13: Investment Risk Modelling

The model

Costs and

benefits

Success or

failure

Losses through inefficie

ncies

Time preference adjustmen

t (discounti

ng)

OutcomeOutcome

OutcomeOutcome

Outcome

Overall project outcome

Overall project outcome

Overall project outcome

Overall project outcome

Overall project outcome

Multiple stakeholders

Monte Carlo simulation10,000 plausible outcomes

Separate runs including and

excluding political risk

factors

Model structure

Page 14: Investment Risk Modelling

Results

Outcomes for Wajir … considering all risks

77.3% of project failure or unprofitability!

Probably too risky

Can risks be mitigated?

-1000 -500 0 500 1000Net present benefits

(million USD)

Pro

bab

ilit

yLo

w

h

igh

Chance of gains: 22.7%Chance of losses: 36.8%

Chance of zero: 40.6%

Page 15: Investment Risk Modelling

Results

Outcomes for Wajir … considering all risks

Variable Importance – where are the

important knowledge

gaps?

Future water purchase by person

Future water priceValue of surviving infant

Additional surviving infantsValue of disease treatment

Water revenue trendNumber of disease treatments

savedLosses due to poor project

designInadequate benefit sharing

Political interferenceLack of cooperation

Risk of increased salinityRisk of drying wells

Page 16: Investment Risk Modelling

Results

Outcomes for Wajir … considering all risks

Future water purchase by person

Future water priceValue of surviving infant

Additional surviving infantsValue of disease treatment

Water revenue trendNumber of disease treatments

savedLosses due to poor project

designInadequate benefit sharing

Political interferenceLack of cooperation

Risk of increased salinityRisk of drying wells

Variable Importance – where are the

important knowledge

gaps?

Can these risks be eliminated?

Page 17: Investment Risk Modelling

-1500 -1000 -500 0 500Net present benefits

(million USD)

Results

Outcomes for Wajir … excluding political risks

Chance of gains: 41.3%Chance of losses: 52.6%

Chance of zero: 6.1%

Pro

bab

ilit

yLo

w

h

igh

Page 18: Investment Risk Modelling

Results

Outcomes for Habaswein … excluding political risks

Chance of losses: 10.9%

Chance of zero: 6.1%

Better prospects than for Wajir

Lower expected water price Payments for ecosystem services

-100 0 100 200 300 400Net present benefits

(million USD)

Pro

bab

ilit

yLo

w

h

igh

Chance of gains: 83.0%

Page 19: Investment Risk Modelling

Net present benefits (million USD)

Results

The water company … excluding political risks

Chance of gains: 78.0%Losses: 22.0%

Good chance of substantial gains

But risk of loss not negligible

0 500 1000 1500 2000

Pro

bab

ilit

yLo

w

h

igh

Page 20: Investment Risk Modelling

Net present benefits (million USD)

Results

Overall outcomes … excluding political risks

Chance of gains: 72.4%Chance of losses:

27.6%

Good prospects for net benefits

Also substantial chance of loss

0 500 1000 1500

Pro

bab

ilit

yLo

w

h

igh

Page 21: Investment Risk Modelling

Results

Overall outcomes … with all risks

Very risky

More likely to fail than to succeed

-500 0 500 1000Net present benefits

(million USD)

Pro

bab

ilit

yLo

w

h

igh

Chance of gains: 37.9%Chance of losses: 62.1%

Page 22: Investment Risk Modelling

Results … all risks considered

downstream

upstream Pipeline communities

donor

Probably nil; losses more likely than

gains

Probably nil; gains more likely than

losses

Gains or losses; risky

Certain loss (or

nil)

Probably nil; gains more likely than

losses

Certain loss

Gains or losses; risky

Water company

Page 23: Investment Risk Modelling

Results … no political obstacles

downstream

upstream Pipeline communities

donorGains or

losses; risky

Gains more likely than

losses

Gains more likely than

losses

Certain loss (or nil)

Gains more likely than

losses

Certain loss

Likely profitable,

but still risky

Water company

Page 24: Investment Risk Modelling

Main sources of uncertainty

Disease treatment

The life of an infant (currently 200 to 100,000 USD)

Valuations

# of saved disease treatment

# of additional surviving infants

Impacts

Water users and water use per user

Price of water

Markets

Political

Hydrological and environmental

Risks

Page 25: Investment Risk Modelling

Risks

Environmental impacts

Risk of salinity, thickness of freshwater layer

Hydro, environment

Careful site selection (deep saline layer, far from Habaswein wells)

Groundwater recharge

Careful borehole design, interception wells

Risk mitigation

Political interference

Inadequate benefit sharing

Suboptimal cooperation

Political

Stakeholder inclusion

Benefit transfer mechanisms

Extra efforts to reduce hydrological risks

Risk mitigation

Page 26: Investment Risk Modelling

Conclusions

Given the current state of uncertainty (of the estimators), the project is very risky, for all stakeholders and as a business

Targeted measurements would help firm up the business case

Major uncertainties are related to stakeholder inclusion and benefit sharing

Hydrologically, the risk of salinity intrusion seemed greater than that of wells running dry

Careful project design should be able to minimize the hydrological risk

Page 27: Investment Risk Modelling

Recommendations

Test drills! Depth to saline water?

Careful cost/benefit analysis for the water company

Value of a life?

Market study to determine likely water price

Revisit of the present modeling approach with more concrete information on project design

Measurement Stakeholder inclusion

Benefit transfer (e.g. PES)

Select sites and optimize borehole technology to minimize risks

Careful aquifer monitoring for early detection of problems

Invest in robust technology to minimize inefficiencies and ensure reliable operation

Design / implementation

Page 28: Investment Risk Modelling

Take-home message

Targeted measurements, careful project design and a dedicated

monitoring system will be essential for making this project a success!

Page 29: Investment Risk Modelling

Questions, comments?

[email protected]

?Image source: http://www.clintonhealth.org

Page 30: Investment Risk Modelling

Water supply for Wajir

http://www.mercycorps.org

Page 31: Investment Risk Modelling

Water supply for Wajir

http://www.mercycorps.org

City of Wajir has major water problems: water supply and sanitation

What are the benefits? What are the costs?

Is it worth doing?

Water pipeline from Habaswein could help

Page 32: Investment Risk Modelling

Stakeholders

downstream

upstream Water company

donor

excited?

apprehensive?

unclear?

concerned?

impatient?

generous?

Page 33: Investment Risk Modelling

Stakeholders

Benefits

Costs

Benefits

Costs

Benefits

Costs

upstream

Benefits

Costs

downstream

Benefits

Costs

Water company

Benefits

Costs

donor

Risks to the projec

t

Page 34: Investment Risk Modelling

http://www.mercycorps.org

The modeling process

Participatory modeling

Calibration training for stakeholdersSolicitation of

estimates

Model runs with stakeholder estimates

Identification of high-value variables

Measurements

Model refinement

Identification of uncertainties

Page 35: Investment Risk Modelling

Stakeholder consultation

http://www.mercycorps.org

1-day stakeholder workshop Around 30 participants, many from Wajir county, including

senator 2-day core groups modeling workshop First draft of model developed Currently 132 uncertain variables

Page 36: Investment Risk Modelling

Costs and benefits for Wajir

http://www.mercycorps.org

CostsPayments for water

Payments for ecosystem servicesInfrastructure maintenance

Etc.

BenefitsPublic health benefits:Reduced infant mortalityLower treatment costs

Sanitation benefitsHigher productivity

Attractive location benefits:Higher investment

Job creationHigher taxes

More demand for local productsReduced brain drain

Clean reliable water:Reduced reliance on wellsWater during dry season

General livelihood improvement

Page 37: Investment Risk Modelling

Risks to the project

http://www.mercycorps.org

RisksNegative feasibility report

Low water yieldInadequate benefit sharing

Political interferenceDrying wells

Increased salinityMaintenance problems

Pipeline failureWater too expensive

Regional conflict

Page 38: Investment Risk Modelling

The model – overview

http://www.mercycorps.org

Page 39: Investment Risk Modelling

The model – benefit calculation

http://www.mercycorps.org

Page 40: Investment Risk Modelling

The model – cost calculation

http://www.mercycorps.org

Page 41: Investment Risk Modelling

The model – risk adjustment

http://www.mercycorps.org

Page 42: Investment Risk Modelling

Preliminary results

Bottom line for Habaswein

Modeled benefits for different stakeholders

Bottom line for Wajir

Page 43: Investment Risk Modelling

Preliminary results

Overall project bottom line

Modeled bottom line for the entire project

Page 44: Investment Risk Modelling

Preliminary results

Important variables(determined by PLS regression)

Page 45: Investment Risk Modelling

Preliminary results

Important variables(determined by PLS regression)

Page 46: Investment Risk Modelling

Preliminary results

High-value variables:

Knowledge gaps that should be closed for better decision making

Random factors that affect decision outcomes

Entry-points for modifications to the project design that increase chances of success

For Wajir project, analysis suggests that more stakeholder involvement will substantially reduce risks

Page 47: Investment Risk Modelling

Lessons learned

Most decisions can be described with fairly simple models

Estimating the range of probable outcomes from a decision is normally possible

Information gaps are often smaller than it seems

Important variables are often unexpected and can be easy to measure

Including stakeholders in model-building is possible and enhances buy-in

Page 48: Investment Risk Modelling

Thanks for your attention! e.luedeling@cgiar.

org