VOI: value of information
VOI: Value of Information (or Very Opaque Inferences?) Pete Naylor
SPE Evening Meeting, 27 November 2012
Legal notice Certain statements included in this presentation contain forward-looking information concerning BG Group plc‟s strategy, operations, financial performance or condition, outlook, growth opportunities or circumstances in the countries, sectors or markets in which BG Group plc operates. By their nature, forward-looking statements involve uncertainty because they depend on future circumstances, and relate to events, not all of which are within BG Group plc‟s control or can be predicted by BG Group plc. Although BG Group plc believes that the expectations reflected in such forward-looking statements are reasonable, no assurance can be given that such expectations will prove to have been correct. Actual results could differ materially from the guidance given in this presentation for a number of reasons. For a detailed analysis of the factors that may affect our business, financial performance or results of operations, we urge you to look at the “Principal risks and uncertainties” included in the BG Group plc Annual Report & Accounts 2011. Nothing in this presentation should be construed as a profit forecast and no part of this presentation constitutes, or shall be taken to constitute, an invitation or inducement to invest in BG Group plc or any other entity, and must not be relied upon in any way in connection with any investment decision. BG Group plc undertakes no obligation to update any forward-looking statements. No representation or warranty, express or implied, is or will be made in relation to the accuracy or completeness of the information in this presentation and no responsibility or liability is or will be accepted by BG Group plc or any of its respective subsidiaries, affiliates and associated companies (or by any of their respective officers, employees or agents) in relation to it.
2
Aims of this presentation To provide an understanding of „value of information‟ (VOI) analysis • When?
• Why? • How?
3
Aims of this presentation To provide an understanding of „value of information‟ (VOI) analysis • When?
• Why? • How?
4
When might VOI analysis be valuable? • Facing a number of decisions • Outcomes are uncertain • Opportunity to acquire additional information
• Information costs money or time
Is the additional information worth the cost?
5
Why might VOI analysis be valuable? • The additional information might reduce future uncertainties • Decisions might change in the light of the new information
If the VOI analysis indicates that no decision changes, then do not waste the money & time in acquiring the new information
6
Key questions • How much does the information cost? – Acquisition – Analysis
– Delay to development
• How reliable is the information? – Will the measurement fail? – False results (imperfect information)
• How useful is the information? – How significant is the parameter(s) to be measured? – What difference will the information make?
VOI analysis can address these questions before investing in the information
7
How do I undertake a VOI analysis? • Case example – Should an appraisal well be drilled in the North Extension? – Should the North Extension be developed?
• A new user took < two hours to learn the software & complete this analysis
8
Influence diagram Reserves = COS * GIIP * RF NPV = Reserves*(NPV/boe) - Costs
Reserves = COS * GIIP * RF
KEY Decision
Chance
Value
Actual COS
Measured COS Sanction development?
Initial decision
Reserves
NPV
Actual GIIP
Measured GIIP
Appraisal cost
RF
Development cost
NPV per boe
A useful thinking tool to assemble the „components‟ of the problem
9
Decision tree
Initial decision
Measured COS
Measured GIIP
Sanction development?
Actual COS
Low Yes
Appraise_now
Yes Nominal
No
No
Actual GIIP
RF
Low
Low
Nominal
Nominal
Yes a
High
NPV,Reserves,PI,Capex
No
NPV,Reserves,PI,Capex High
High NPV,Reserves,PI,Capex
Sanction development?
Yes Develop_now No
a
Sanction development?
Yes Walk_away No
a
Abbreviated form of decision tree, full tree contains 2x3x3 = 18 branches
10
Base case: reserves for North Extension 100% 90%
Probability that reserves less than...
80% 70% 60% 50%
Develop_now
40%
Deterministic = 78MMboe
30% 20% 10% 0% 0
10
20
30
40
50
60
70
80
90
100
110
Reserves (MMboe)
Beware that this is not the basis of your business case
11
Risk profile: reserves for North Extension 100%
Probability of not finding hydrocarbons is
90%
Probability that reserves less than...
80%
(1 – 0.65) = 0.35
70% 60% 50%
Develop_now
40%
Expected value = 40MMboe
30% 20% 10% 0% 0
10
20
30
40
50
60
70
80
90
100
110
Reserves (MMboe)
Illustrates our best estimate of the range of possible reserves
12
Influence diagram Reserves = COS * GIIP * RF NPV = Reserves*(NPV/boe) - Costs
Reserves = COS * GIIP * RF
KEY Decision
Chance
Value
Actual COS
Measured COS Sanction development?
Initial decision
Reserves
NPV
Actual GIIP
Measured GIIP
Appraisal cost
RF
Development cost
NPV per boe
A useful thinking tool to assemble the „components‟ of the problem
13
Influence diagram: extended to include Reserves = COS * GIIP * RF NPV = Reserves*(NPV/boe) - Costs appraisal Reserves = COS * GIIP * RF
KEY Decision
Chance
Value
Actual COS
Measured COS Sanction development?
Initial decision
Reserves
NPV
Actual GIIP
Measured GIIP
Appraisal cost
RF
Development cost
NPV per boe
The measured information depends on the actual state of nature
14
Conditional probabilities: chance of success perfect information
Actual Actual COS Actual COS COS
Yes Yes .65 Yes No .65 .35 No No .35
Measured COS
1 Measured COS
1 0 0
Captures the reliability of the measurement
Yes 1 No 0 Yes 0 No 1
15
Conditional probabilities: chance of success no information
Actual Actual COS Actual COS COS
Yes Yes .65 Yes No .65 .35 No No .35
Measured COS
1 Measured COS
1 0 0
Captures the reliability of the measurement
Yes .65 No .35 Yes .65 No .35
16
Conditional probabilities: chance of success imperfect information
Actual Actual COS Actual COS COS
Yes Yes .65 Yes No .65 .35 No No .35
Measured COS
1 Measured COS
1 0 0
Captures the reliability of the measurement
Yes .8 No .2 Yes .1 No .9
17
Influence diagram: extended to include Reserves = COS * GIIP * RF NPV = Reserves*(NPV/boe) - Costs appraisal Reserves = COS * GIIP * RF
KEY Decision
Chance
Value
Actual COS
Measured COS Sanction development?
Initial decision
Reserves
NPV
Actual GIIP
Measured GIIP
Appraisal cost
RF
Development cost
NPV per boe
The measured information depends on the actual state of nature
18
Conditional probabilities: gas initially in place
Actual GIIP
Low
Measured GIIP
Nominal
Measured GIIP
High
Measured GIIP
Captures the reliability of the measurement
Low .8 Nominal .15 High .05 Low .1 Nominal .8 High .1 Low .05 Nominal .15 High .8 19
Decision tree
Initial decision
Measured COS
Measured GIIP
Sanction development?
Actual COS
Low Yes
Appraise_now
Low Yes
Nominal No
Actual GIIP
No
RF
Low
Yes a
High
NPV,Reserves,PI,Capex Nominal
Nominal
High
High
No
NPV,Reserves,PI,Capex NPV,Reserves,PI,Capex
Sanction development?
Yes Develop_now No
a
Sanction development?
Yes Walk_away No
a
Abbreviated form of decision tree, full tree contains 2x3x3 = 18 branches
20
Decision tree: extended to include appraisal Initial decision
Measured COS
Measured GIIP
Sanction development?
Actual COS
Low Yes
Appraise_now
Yes Nominal
No
No
Actual GIIP
RF
Low
Low
Nominal
Nominal
High
High
Yes a
High
NPV,Reserves,PI,Capex
No
NPV,Reserves,PI,Capex NPV,Reserves,PI,Capex
Sanction development?
Yes Develop_now No
a
Sanction development?
Yes Walk_away No
Captures the time order of events
a
21
Influence diagram Reserves = COS * GIIP * RF NPV = Reserves*(NPV/boe) - Costs
Reserves = COS * GIIP * RF
KEY Decision
Chance
Value
Actual COS
Measured COS Sanction development?
Initial decision
Reserves
NPV
Actual GIIP
Measured GIIP
Appraisal cost
RF
Development cost
NPV per boe
The measured information depends on the actual state of nature
22
Influence diagram: extended to include Reserves = COS * GIIP * RF economics NPV = Reserves*(NPV/boe) - Costs
Reserves = COS * GIIP * RF
KEY Decision
Chance
Actual COS
Measured COS Sanction development?
Initial decision
RF
Reserves
NPV
Actual GIIP
Measured GIIP
Appraisal cost
NPV = Reserves*(NPV/boe) - Capex
Value
Development cost
Which decision options give the best NPV?
NPV per boe 23
NPV risk profile: develop now 100% 90%
Probability that NPV less than...
80% 70%
Develop @ $300MM & find no hydrocarbons
60%
Appraise_now
50% Develop_now
40%
Expected value = $60MM
30% 20% 10% 0% -300
-200
-100
0
100
200
300
400
500
600
700
800
Net Present Value
Illustrates the range of possible outcomes
24
NPV risk profile: develop now 100% 90%
Probability that NPV less than...
80% 70% 60%
Appraise_now
50% Develop_now
40% 30% 20% 10% 0% -300
-200
-100
0
100
200
300
400
500
600
700
800
Net Present Value
Illustrates the range of possible outcomes
25
NPV risk profile: appraise, no information 100% 90%
Probability that NPV less than...
80% 70% 60%
Appraise_now
50% Develop_now
40% 30%
NPV decreased by $40MM
20% 10% 0% -300
-200
-100
0
100
200
300
400
500
600
700
800
Net Present Value
Spend the money on appraisal but get no information
26
NPV risk profile: appraise, perfect information 100% 90%
Probability that NPV less than...
80% 70% 60%
Appraise_now
50% Develop_now
40% 30% 20% 10% 0% -300
-200
-100
0
100
200
300
400
500
600
700
800
Net Present Value ($MM)
Perfect information means field is only developed when there is a net gain
27
NPV risk profile: appraise, imperfect information 100% 90%
Probability that NPV less than...
80% 70% 60%
Appraise_now
50% Develop_now
40% 30% 20% 10% 0% -300
-200
-100
0
100
200
300
400
500
600
700
800
Net Present Value ($MM)
Imperfect information means that sometimes you get it wrong
28
Policy tree: identifies the optimum decisions
Yes
Actual COS [-50]
Sanction develop- ment? Low
[-40] Actual COS
30% No
[-40]
Actual COS Yes Measured GIIP Yes Measured COS Appraise_now
[181]
56%
[206]
Sanction develop- ment? Nominal
[206] Actual COS
41%
[83]
No
Initial decision
[-40]
Actual COS
[83]
Yes
[368]
Sanction develop- ment? High
[368] Actual COS
30% No
[-40]
Measured GIIP No
[-40]
44% Actual COS Develop_now
[60]
Actual COS Walk_away
[-0]
Note the appraisal information alters the development decision
29
NPV risk profile: appraise, imperfect information 100% 90% 80%
Probability that NPV less than...
70% 60%
Appraise_now
50% Develop_now
40% 30% 20% 10% 0% -300
-200
-100
0
100 82.7 59.6
200
300
Net Present Value ($MM)
400
500
600
700
30
Two-way rainbow diagram 120 110 100
Develop now
Walk away
Appraisal cost ($MM)
90 80 70 60 50 40 30 20
Appraise now
10 0 100
180
260
340
420
500
580
660
740
Development cost ($MM)
Green indicates where appraisal give highest expected NPV
31
NPV risk profile: appraise, imperfect information 100% 90% 80%
Probability that NPV less than...
70% 60%
Appraise_now
50% Develop_now
40% 30% 20% 10% 0% -300
-200
-100
0
100 82.7 59.6
200
300
Net Present Value ($MM)
400
500
600
700
32
Risk appetite: which choice would you make if appraisal cost $80MM? 100% 90% 80%
Probability that NPV less than...
70% 60%
Appraise_now
50% Develop_now
How much expected NPV would you sacrifice to avoid the downside?
40% 30% 20% 10% 0% -400
-300
-200
-100
0
100
200
300
400
500
600
700
800
42.7 59.6
Net Present Value ($MM)
33
Summary: Value of information • When? – Facing a number of decisions – Outcomes are uncertain
– Opportunity to acquire additional information – Information costs money or time
• Why? – The additional information might reduce future uncertainties – Decisions might change in the light of the new information
• How? – Invest two hours and get a Decision Analyst to show you how
34
“What the hell is always the right decision!” Marilyn Monroe (1926-1962)
May have worked for Marilyn, but your management will require more analysis
35
SPE Evening Meeting, 27 November 2012
Legal notice Certain statements included in this presentation contain forward-looking information concerning BG Group plc‟s strategy, operations, financial performance or condition, outlook, growth opportunities or circumstances in the countries, sectors or markets in which BG Group plc operates. By their nature, forward-looking statements involve uncertainty because they depend on future circumstances, and relate to events, not all of which are within BG Group plc‟s control or can be predicted by BG Group plc. Although BG Group plc believes that the expectations reflected in such forward-looking statements are reasonable, no assurance can be given that such expectations will prove to have been correct. Actual results could differ materially from the guidance given in this presentation for a number of reasons. For a detailed analysis of the factors that may affect our business, financial performance or results of operations, we urge you to look at the “Principal risks and uncertainties” included in the BG Group plc Annual Report & Accounts 2011. Nothing in this presentation should be construed as a profit forecast and no part of this presentation constitutes, or shall be taken to constitute, an invitation or inducement to invest in BG Group plc or any other entity, and must not be relied upon in any way in connection with any investment decision. BG Group plc undertakes no obligation to update any forward-looking statements. No representation or warranty, express or implied, is or will be made in relation to the accuracy or completeness of the information in this presentation and no responsibility or liability is or will be accepted by BG Group plc or any of its respective subsidiaries, affiliates and associated companies (or by any of their respective officers, employees or agents) in relation to it.
2
Aims of this presentation To provide an understanding of „value of information‟ (VOI) analysis • When?
• Why? • How?
3
Aims of this presentation To provide an understanding of „value of information‟ (VOI) analysis • When?
• Why? • How?
4
When might VOI analysis be valuable? • Facing a number of decisions • Outcomes are uncertain • Opportunity to acquire additional information
• Information costs money or time
Is the additional information worth the cost?
5
Why might VOI analysis be valuable? • The additional information might reduce future uncertainties • Decisions might change in the light of the new information
If the VOI analysis indicates that no decision changes, then do not waste the money & time in acquiring the new information
6
Key questions • How much does the information cost? – Acquisition – Analysis
– Delay to development
• How reliable is the information? – Will the measurement fail? – False results (imperfect information)
• How useful is the information? – How significant is the parameter(s) to be measured? – What difference will the information make?
VOI analysis can address these questions before investing in the information
7
How do I undertake a VOI analysis? • Case example – Should an appraisal well be drilled in the North Extension? – Should the North Extension be developed?
• A new user took < two hours to learn the software & complete this analysis
8
Influence diagram Reserves = COS * GIIP * RF NPV = Reserves*(NPV/boe) - Costs
Reserves = COS * GIIP * RF
KEY Decision
Chance
Value
Actual COS
Measured COS Sanction development?
Initial decision
Reserves
NPV
Actual GIIP
Measured GIIP
Appraisal cost
RF
Development cost
NPV per boe
A useful thinking tool to assemble the „components‟ of the problem
9
Decision tree
Initial decision
Measured COS
Measured GIIP
Sanction development?
Actual COS
Low Yes
Appraise_now
Yes Nominal
No
No
Actual GIIP
RF
Low
Low
Nominal
Nominal
Yes a
High
NPV,Reserves,PI,Capex
No
NPV,Reserves,PI,Capex High
High NPV,Reserves,PI,Capex
Sanction development?
Yes Develop_now No
a
Sanction development?
Yes Walk_away No
a
Abbreviated form of decision tree, full tree contains 2x3x3 = 18 branches
10
Base case: reserves for North Extension 100% 90%
Probability that reserves less than...
80% 70% 60% 50%
Develop_now
40%
Deterministic = 78MMboe
30% 20% 10% 0% 0
10
20
30
40
50
60
70
80
90
100
110
Reserves (MMboe)
Beware that this is not the basis of your business case
11
Risk profile: reserves for North Extension 100%
Probability of not finding hydrocarbons is
90%
Probability that reserves less than...
80%
(1 – 0.65) = 0.35
70% 60% 50%
Develop_now
40%
Expected value = 40MMboe
30% 20% 10% 0% 0
10
20
30
40
50
60
70
80
90
100
110
Reserves (MMboe)
Illustrates our best estimate of the range of possible reserves
12
Influence diagram Reserves = COS * GIIP * RF NPV = Reserves*(NPV/boe) - Costs
Reserves = COS * GIIP * RF
KEY Decision
Chance
Value
Actual COS
Measured COS Sanction development?
Initial decision
Reserves
NPV
Actual GIIP
Measured GIIP
Appraisal cost
RF
Development cost
NPV per boe
A useful thinking tool to assemble the „components‟ of the problem
13
Influence diagram: extended to include Reserves = COS * GIIP * RF NPV = Reserves*(NPV/boe) - Costs appraisal Reserves = COS * GIIP * RF
KEY Decision
Chance
Value
Actual COS
Measured COS Sanction development?
Initial decision
Reserves
NPV
Actual GIIP
Measured GIIP
Appraisal cost
RF
Development cost
NPV per boe
The measured information depends on the actual state of nature
14
Conditional probabilities: chance of success perfect information
Actual Actual COS Actual COS COS
Yes Yes .65 Yes No .65 .35 No No .35
Measured COS
1 Measured COS
1 0 0
Captures the reliability of the measurement
Yes 1 No 0 Yes 0 No 1
15
Conditional probabilities: chance of success no information
Actual Actual COS Actual COS COS
Yes Yes .65 Yes No .65 .35 No No .35
Measured COS
1 Measured COS
1 0 0
Captures the reliability of the measurement
Yes .65 No .35 Yes .65 No .35
16
Conditional probabilities: chance of success imperfect information
Actual Actual COS Actual COS COS
Yes Yes .65 Yes No .65 .35 No No .35
Measured COS
1 Measured COS
1 0 0
Captures the reliability of the measurement
Yes .8 No .2 Yes .1 No .9
17
Influence diagram: extended to include Reserves = COS * GIIP * RF NPV = Reserves*(NPV/boe) - Costs appraisal Reserves = COS * GIIP * RF
KEY Decision
Chance
Value
Actual COS
Measured COS Sanction development?
Initial decision
Reserves
NPV
Actual GIIP
Measured GIIP
Appraisal cost
RF
Development cost
NPV per boe
The measured information depends on the actual state of nature
18
Conditional probabilities: gas initially in place
Actual GIIP
Low
Measured GIIP
Nominal
Measured GIIP
High
Measured GIIP
Captures the reliability of the measurement
Low .8 Nominal .15 High .05 Low .1 Nominal .8 High .1 Low .05 Nominal .15 High .8 19
Decision tree
Initial decision
Measured COS
Measured GIIP
Sanction development?
Actual COS
Low Yes
Appraise_now
Low Yes
Nominal No
Actual GIIP
No
RF
Low
Yes a
High
NPV,Reserves,PI,Capex Nominal
Nominal
High
High
No
NPV,Reserves,PI,Capex NPV,Reserves,PI,Capex
Sanction development?
Yes Develop_now No
a
Sanction development?
Yes Walk_away No
a
Abbreviated form of decision tree, full tree contains 2x3x3 = 18 branches
20
Decision tree: extended to include appraisal Initial decision
Measured COS
Measured GIIP
Sanction development?
Actual COS
Low Yes
Appraise_now
Yes Nominal
No
No
Actual GIIP
RF
Low
Low
Nominal
Nominal
High
High
Yes a
High
NPV,Reserves,PI,Capex
No
NPV,Reserves,PI,Capex NPV,Reserves,PI,Capex
Sanction development?
Yes Develop_now No
a
Sanction development?
Yes Walk_away No
Captures the time order of events
a
21
Influence diagram Reserves = COS * GIIP * RF NPV = Reserves*(NPV/boe) - Costs
Reserves = COS * GIIP * RF
KEY Decision
Chance
Value
Actual COS
Measured COS Sanction development?
Initial decision
Reserves
NPV
Actual GIIP
Measured GIIP
Appraisal cost
RF
Development cost
NPV per boe
The measured information depends on the actual state of nature
22
Influence diagram: extended to include Reserves = COS * GIIP * RF economics NPV = Reserves*(NPV/boe) - Costs
Reserves = COS * GIIP * RF
KEY Decision
Chance
Actual COS
Measured COS Sanction development?
Initial decision
RF
Reserves
NPV
Actual GIIP
Measured GIIP
Appraisal cost
NPV = Reserves*(NPV/boe) - Capex
Value
Development cost
Which decision options give the best NPV?
NPV per boe 23
NPV risk profile: develop now 100% 90%
Probability that NPV less than...
80% 70%
Develop @ $300MM & find no hydrocarbons
60%
Appraise_now
50% Develop_now
40%
Expected value = $60MM
30% 20% 10% 0% -300
-200
-100
0
100
200
300
400
500
600
700
800
Net Present Value
Illustrates the range of possible outcomes
24
NPV risk profile: develop now 100% 90%
Probability that NPV less than...
80% 70% 60%
Appraise_now
50% Develop_now
40% 30% 20% 10% 0% -300
-200
-100
0
100
200
300
400
500
600
700
800
Net Present Value
Illustrates the range of possible outcomes
25
NPV risk profile: appraise, no information 100% 90%
Probability that NPV less than...
80% 70% 60%
Appraise_now
50% Develop_now
40% 30%
NPV decreased by $40MM
20% 10% 0% -300
-200
-100
0
100
200
300
400
500
600
700
800
Net Present Value
Spend the money on appraisal but get no information
26
NPV risk profile: appraise, perfect information 100% 90%
Probability that NPV less than...
80% 70% 60%
Appraise_now
50% Develop_now
40% 30% 20% 10% 0% -300
-200
-100
0
100
200
300
400
500
600
700
800
Net Present Value ($MM)
Perfect information means field is only developed when there is a net gain
27
NPV risk profile: appraise, imperfect information 100% 90%
Probability that NPV less than...
80% 70% 60%
Appraise_now
50% Develop_now
40% 30% 20% 10% 0% -300
-200
-100
0
100
200
300
400
500
600
700
800
Net Present Value ($MM)
Imperfect information means that sometimes you get it wrong
28
Policy tree: identifies the optimum decisions
Yes
Actual COS [-50]
Sanction develop- ment? Low
[-40] Actual COS
30% No
[-40]
Actual COS Yes Measured GIIP Yes Measured COS Appraise_now
[181]
56%
[206]
Sanction develop- ment? Nominal
[206] Actual COS
41%
[83]
No
Initial decision
[-40]
Actual COS
[83]
Yes
[368]
Sanction develop- ment? High
[368] Actual COS
30% No
[-40]
Measured GIIP No
[-40]
44% Actual COS Develop_now
[60]
Actual COS Walk_away
[-0]
Note the appraisal information alters the development decision
29
NPV risk profile: appraise, imperfect information 100% 90% 80%
Probability that NPV less than...
70% 60%
Appraise_now
50% Develop_now
40% 30% 20% 10% 0% -300
-200
-100
0
100 82.7 59.6
200
300
Net Present Value ($MM)
400
500
600
700
30
Two-way rainbow diagram 120 110 100
Develop now
Walk away
Appraisal cost ($MM)
90 80 70 60 50 40 30 20
Appraise now
10 0 100
180
260
340
420
500
580
660
740
Development cost ($MM)
Green indicates where appraisal give highest expected NPV
31
NPV risk profile: appraise, imperfect information 100% 90% 80%
Probability that NPV less than...
70% 60%
Appraise_now
50% Develop_now
40% 30% 20% 10% 0% -300
-200
-100
0
100 82.7 59.6
200
300
Net Present Value ($MM)
400
500
600
700
32
Risk appetite: which choice would you make if appraisal cost $80MM? 100% 90% 80%
Probability that NPV less than...
70% 60%
Appraise_now
50% Develop_now
How much expected NPV would you sacrifice to avoid the downside?
40% 30% 20% 10% 0% -400
-300
-200
-100
0
100
200
300
400
500
600
700
800
42.7 59.6
Net Present Value ($MM)
33
Summary: Value of information • When? – Facing a number of decisions – Outcomes are uncertain
– Opportunity to acquire additional information – Information costs money or time
• Why? – The additional information might reduce future uncertainties – Decisions might change in the light of the new information
• How? – Invest two hours and get a Decision Analyst to show you how
34
“What the hell is always the right decision!” Marilyn Monroe (1926-1962)
May have worked for Marilyn, but your management will require more analysis
35
