Incorporating Risk in Strategic Decision Making Processes S T R A T E G I C INFORMATION CONSULTANTS Donna Hill Prudence Thompson Emma Smith Dr John Henstridge Data Analysis Australia Overview Uncertainty of predicted values Ignored in many instances Equates to risk in many instances Case studies Agricultural investment company Modelling uncertainty in expected value of a crop Improvement strategies Reviewed client’s modelling of risk Simulation approach to estimate uncertainty STRATEGIC INFORMATION CONSULTANTS. www.daa.com.au Page 2 Uncertainty Provide clients with an estimated or predicted value Based on models Based on survey results Based on simulations Uncertainty of that value We can explain that there is uncertainty We can quantify the uncertainty STRATEGIC INFORMATION CONSULTANTS. www.daa.com.au Page 3 Case 1 – The Problem Agricultural investment company Can estimate expected yield Historic data Can get predictions of price Forward pricing on stock market Hence can get predictions of value Yield times price Investor return based on crop value and fees charged What fees should be charged? Need to use uncertainty in predicted values STRATEGIC INFORMATION CONSULTANTS. www.daa.com.au Page 4 Case 1 – Our Approach Understand the distributions Of historic yield data At farmer level At shire level At state level Of pricing data Forward prices What the market thinks the grain is going to be worth Final prices What the grain actually sold for STRATEGIC INFORMATION CONSULTANTS. www.daa.com.au Page 5 0.7 Historic Yield Distributions 0.4 0.3 0.0 0.1 0.2 Density 0.5 0.6 2001 2002 2003 2004 2005 2006 2007 2008 0 STRATEGIC INFORMATION CONSULTANTS. www.daa.com.au 1 2 3 Yield (g/t) 4 5 6 Page 6 300 250 150 200 Final Price ($) 350 400 Historic Pricing Data 50 100 150 200 250 300 350 Forward Price ($) STRATEGIC INFORMATION CONSULTANTS. www.daa.com.au Page 7 0.010 Forward Price Final Price 0.000 0.005 Density 0.015 Historic Pricing Distributions 100 STRATEGIC INFORMATION CONSULTANTS. www.daa.com.au 150 200 Price ($) 250 Page 8 Case 1 – Our Approach (cont.) Appropriately combine models for yield and pricing Leads to understanding of the distribution of crop values Obtain predictions for crop value And uncertainty Calculate the distribution for investor return For a given fee structure Expected value and probability of achieving a particular investor return Customise fee structure to give high probability of achieving desired return STRATEGIC INFORMATION CONSULTANTS. www.daa.com.au Page 9 Case 1 – The Result Provided client with modelling tool User sets desired investor return User sets desired probability of achieving this Tool models distributions at farmer level Tool sets appropriate costs at farmer level Distribution of returns for an individual To obtain a portfolio level return Client can test different scenarios STRATEGIC INFORMATION CONSULTANTS. By changing many of the inputs www.daa.com.au Page 10 Case 2 – The Problem Client wanted to implement a set of improvement strategies Needed to model the impact Predict benefits Incorporating and combining uncertainty of each strategy Justify expenditure Monte Carlo approach Wanted Data Analysis Australia to review methodology and results STRATEGIC INFORMATION CONSULTANTS. www.daa.com.au Page 11 Case 2 – The Review Identify major areas of uncertainty Four areas Review model assumptions Assumes a triangular probability density function for each improvement strategy STRATEGIC INFORMATION CONSULTANTS. worst case scenario www.daa.com.au most likely scenario best case scenario Page 12 Case 2 – The Result Suggested a simpler, more transparent approach Sum means and variances Same results as Monte Carlo approach Recommended an approach to evaluate future improvement strategies Expand on the Monte Carlo approach Include other areas of uncertainty Incorporate different distributions STRATEGIC INFORMATION CONSULTANTS. www.daa.com.au Page 13 Case 3 – The Problem Client needed to determine appropriate pricing structure for future contracts Client’s costs affected by overall peak demand for their product Higher demand means higher costs to meet that demand Customer’s contract based on how much their usage is expected to add to overall peak Key questions What will the overall peak demand be? Can model this based on historic data How much will a customer contribute to that peak? Use last year’s data for that customer to predict What if last year’s data was ‘unusual’? Usage often influenced by weather STRATEGIC INFORMATION CONSULTANTS. www.daa.com.au Page 14 35 30 25 15 20 Maximum Temp 40 45 Historic Temperature Data 0 STRATEGIC INFORMATION CONSULTANTS. www.daa.com.au 20 40 Index 60 80 Page 15 0.08 Historic Temperature Distributions 0.04 0.00 0.02 Density 0.06 2002 2003 2004 2005 2006 2007 2008 2009 15 STRATEGIC INFORMATION CONSULTANTS. www.daa.com.au 20 25 30 35 Maximum Temp 40 45 50 Page 16 Case 3 – Our Approach Modelling and simulation approach Model overall demand Several years’ data available Model individual customer usage Only one year’s data available Obtain historic weather data Over many years For each year of weather data Calculate fitted values for overall demand Determine peaks for each year Calculate fitted values for customer usage At the peak times for each year Distribution for each customer STRATEGIC INFORMATION CONSULTANTS. Variation in contribution to peaks www.daa.com.au Page 17 Case 3 – The Result Analysed results by customer group Many had high variability in demand High risk in using a single year’s data for setting contract prices Allowed client to understand customer groups better Develop strategies to set more appropriate pricing for future contracts STRATEGIC INFORMATION CONSULTANTS. www.daa.com.au Page 18 Summary Clients do care about uncertainty They realise it represents risk They want to incorporate it into their decision making STRATEGIC INFORMATION CONSULTANTS. www.daa.com.au Page 19 Thank You Donna Hill S T R A T E G I C INFORMATION CONSULTANTS [email protected] www.daa.com.au (08) 9386 3304
© Copyright 2026 Paperzz