Final Exam: Large Instrument Manufacturer, Inc. ARUNABHA SAHA Recommendations Our intial recommendation to LIM at this point is to pursue the alternative that involves training UDC to handle Beta failures. We further recommend a subsequent analysis taking into account much of the information from the first round and assessing the uncertainties. Outline Decision Tools Used Framing Decision Hierarchy Decision Diagram Deterministic Analysis NPV of Alternatives Deterministic Sensitivity Probabilistic Analysis Decision Tree Probabilistic Dominance Appraisal Sensitivity Analysis Value of Information and Control Recommendations for future analysis Decision Hierarchy • Policy is to be Risk-Neutral in this Range. • Making a profit is the end goal of this exercise. Selling a product but not making a profit is not a desirable option. • ROE-II will be sold only for a period of 5 years hence any strategy should be optimized for the next 5 year period. Policy Strategy Tactics To sell ROE-II to UDC or not ? Eliminate or mitigate the risk associated with servicing that caused the bad experience with ROE-II Determine the uncertainties early and how they affect the eventual outcome. Concentrate on evaluating the important uncertainties before making the decision Attempt to control the important uncertainties by means at our disposal Decision Diagram Num of Αlpha Failures Total Alpha Failure Cost Sell ROE-II to UDC Cost of Alpha Failures Num of Beta Failures Cost of Beta Failures Total Beta Failure Cost Units Sold Total Revenue Continue with Existing Arrangement Training Cost Initial Cost (Training + Installation) Train UDC engineers OR Outsource Value NPV of Each Strategy at Base Case 200 178 150 150 NPV in $1000s 100 50 0 Existing Train UDC Contract IPX Do not -8 Sell -50 -100 -105 -150 We can see that in the base case • “Training UDC” is the best option • Contracting with IPX is also an attractive • Existing option is far worse than either training or contracting Comparison of Alternatives (100% range) Range of Profit for the 4 Alternatives Do not Sell Mid: 147K Contract IPX Mid:178K Mid:-105K Train WDC Existing -800 -600 -400 -200 0 Profit in $1000s 200 400 600 800 The Key takeaway here is that the “Training UDC” option is the more profitable option in the base case. However the “Train UDC” option also has a higher range (-88, 595) than Contract IPX(0, 347). This “can” imply more risk associated with the “Training WDC” Option Assessing Important Uncertainties (95%) Existing Arrangement Train UDC Beta Repair Cost Training Cost Num of Beta Failures Units Sold Num of Alpha Failures Num of Alpha Failures Units Sold Avg Cost Alpha Repair Cost -$300.00 -$250.00 -$200.00 -$150.00 -$100.00 -$50.00 $0.00 0.00 50.00 100.00 150.00 From the tornado diagrams above we observe that. • • • • 4 Uncertainties to be assessed for Alternative 1 (Existing arrangement). 3 Uncertainties for Alternative 2 (Training UDC for servicing) 1 Uncertainty for Alternative 3 (Contract with IPX) Alternative 4(Do not Sell) has no uncertainties 200.00 250.00 300.00 350.00 Probabilistic Assessment Discretizing (Equal Areas – Shortcut) Low(0-25%) Mid(25-75%) High(75-100) 8.91 9.95 12.5 6.92 8 9.5 Num Alpha failures (per yr) 11.90% 15.00% 19.00% Num Beta failures (per yr) 64.50% 70.00% 75.50% Units Sold 24.5 30 49.5 Training Cost 3.55 8 12.44 Avg Repair Cost(alpha) Avg Repair Cost (Beta) We divided the continuous probabilities into the range {0-25%, 25-75%, 75-100%}. We then arrived at the the values in the table above as the cumulative mean over the individual ranges. We will use these values to further refine our model. Decision Tree (Condensed) Value Measure U-Value Existing Alternative -116.92 -116.92 195.91 195.91 161.75 161.75 -8.00 -8.00 Train UDC 195.9063 Contract IPX Do not Sell • Alternative 2 (Train UDC) has the best CE and is the initial recommendation if there is no room for further analysis • We do recommend further analysis on uncertainties to assess their impact and devise plans to control them or mitigate their impact Probabilistic Dominance 1.2000 1.0000 0.8000 Train UDC Contract IPX 0.6000 Do not Sell Existing 0.4000 0.2000 -400 -300 -200 0.0000 -100 0 100 Profit in $1000s 200 300 400 500 600 This graph shows the CDF of the Certain Equivalent of 4 alternatives Dominance Contd.. Type Dominance Exists Comments/Insight Deterministic No There is always a point where any Alternative X is greater than any other Alternative Y. Therefore you are never 100% sure that the alternative you select is better than any other for the entire universe of possibilities 1st Order Yes This implies that “Train UDC” is a better alternative than the “Existing” alternative over the entire probability space. We can discard “Existing” option for further analysis for almost all practical cases (See note on slide) “Train UDC” Dominates “Existing Option” 2nd Order Yes “Train UDC” Dominates “Contract with IPX” and “Do not Sell” “Contract IPX” Dominates “Do not Sell” Value of Information Value of Information 35 28.6 30 25 20 15 14.31 10 6.28 5 0 Units Sold Training Cost Num of Alpha Failures • If LIM is able to get information on “Training Cost” in advance then $28.6K is the highest it should be willing to pay • Similarly $14.31K for information on “Units Sold” and $6.28K for Alpha failures Value of Control (1 degree) VoC 151.5 151 150.5 150 149.5 149 148.5 148 147.5 147 146.5 151.09 148.28 Units Sold = 50 VoC Training Cost = $12.44K If LIM can control the value of “Units Sold” to 50 units then $151.09K is the most it should be willing to pay, likewise $148.28K for pegging Training Cost to $12.44K Sensitivity Analysis: Unit Cost Closed/Open Loop Sensitivity Analysis (Units Sold) 400 Contract IPX 350 347 300 289.75 Train UDC 250 Train UDC Open Loop 200 Closed Loop 175.625 150 142.625 100 50 0 Low Base High Sensitivity Analysis (Training Cost) Closed/Open Loop Sensitivty on Training Cost 400 350 344.1875 300 250 Train UDC Contract IPX 200 Open 196.063 196.0625 Closed 161.75 150 Train UDC 100 50 47.3125 0 Low Base High Sensitivity Analysis— Num of Alpha Failures Sensitivity to Num of Alpha Repairs 300 250 248.375 Contract IPX 200 Train UDC 199.313 161.75 150 Train UDC 136.625 Train UDC 100 50 0 Low Base High Closed Open Sensitivity Analysis— Risk Aversion Sensitivity To Risk Aversion 1 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045 CE 0.1 CE 0.01 0.001 Risk Aversion Coefficient (γ) The decision does not change with change in risk aversion and Train UDC is the best option Conclusions Items for a second-round assessment Determine Value of Information and VoC for multiple degrees e.g (Units Sold & Training Cost) Identify tools that can effectively provide Control or Information on the uncertainty parameters and evaluate their effectiveness. Revaluate Risk-Preference Study if options be used as a tool to mitigate risk ? Instead of Contracting with IPX, buy the option to contract with IPX. Evaluate an insurance purchase for uncertainties that cannot be controlled. Appendix FULL DECISION TREE Decision Tree (Train WDC Part 1) Value Measure U-Value 0.25 Alpha Failures Low 0.25 Units Sold Low 249 291 291 248 248 209 209 356 356 312 312 256 256 587 587 515 515 422 422 0.5 Alpha Failures Mid 249 0.25 Alpha Failures High 0.25 Alpha Failures Low 0.25 Training Low 344.188 344.188 0.5 Units Sold Base 309 0.5 Alpha Failures Mid 309 0.25 Alpha Failures High 0.25 Alpha Failures Low 0.25 Units Sold High 509.75 0.5 Alpha Failures Mid 509.75 0.25 Alpha Failures High Decision Tree: Train WDC Part 2 0.25 Alpha Failures Low 0.25 Units Sold Low 143.5 182 182 146 146 100 100 222 222 179 179 122 122 367 367 295 295 202 202 0.5 Alpha Failures Mid 143.5 0.25 Alpha Failures High 0.25 Alpha Failures Low 0.5 Training Base 196.063 196.063 0.5 Units Sold Base 175.5 0.5 Alpha Failures Mid 175.5 0.25 Alpha Failures High 0.25 Alpha Failures Low 0.25 Units Sold High 289.75 0.5 Alpha Failures Mid 289.75 0.25 Alpha Failures High Decision Tree: Train WDC Part 3 0.25 Alpha Failures Low 0.25 Units Sold Low 34.5 73 73 37 37 -9 -9 89 89 46 46 -11 -11 147 147 75 75 -18 -18 0.5 Alpha Failures Mid 34.5 0.25 Alpha Failures High 0.25 Alpha Failures Low 0.25 Training High 47.3125 47.3125 0.5 Units Sold Base 42.5 0.5 Alpha Failures Mid 42.5 0.25 Alpha Failures High 0.25 Alpha Failures Low 0.25 Units Sold High 69.75 0.5 Alpha Failures Mid 69.75 0.25 Alpha Failures High Decision Tree: Contract IPX Value Measure U-Value 0.25 Units Sold Low -15 -15 150 150 347 347 0.5 Units Sold Medium 161.75 0.25 Units Sold High