Hugh Cronin Jim Scheer • How to determine the accuracy of CBA • Sources of error • Distribution of net benefits over time • The Coquihalla Highway CBAs • Conclusion • Tips for dealing with error in CBA Project inception Project implementation Project completion in medias res (during) ex ante (before) ex post (after) Study 1 – ex ante Study 2 – in media res Study 3 – ex post Benefits $394.6 $639.4 $1,307.8 Costs $354.4 $767.6 $913.5 Net benefits $40.2 -$128.2 $394.3 Error Factors contributing to increasing levels of error Note! Omissions (Double counting) • • Impacts excluded (e.g. environmental) Counting benefits in primary and secondary markets • Reduces as projects evolve (knowledge increase) Forecasting • Complex/ changing project • Always wrong! • Human response unpredictable • • Cognitive biases (perception, emotion) • Potential for ex post error remains (counterfactual) Also reduce over time • Shadow price determination • Valuation • Measurement • Inaccurate observation or interpretation • Assumptions usually debatable Reduce over time Perceived as relatively less important than other errors Time Distribution ex ante andex ex post post estimates of net Distribution of exofante and estimates of benefits net benefits f t (NB t ) Net benefits Net benefits Error Factors contributing to increasing Note levels of error Omissions (Double counting) 1. 2. Opportunity cost of land Cost of environmental protection 1. 2. Low impact on result Mitigation construction cost included but no pollution costs etc. Forecasting 1. 2. Future Traffic volumes Time and distance savings (Incl. reduced congestion on alt. routes) Accident rates 1. 2. Refined over time (toll receipts) Range of estimates and methods 3. Range of estimates 1. 2. 3. Value of time savings Valuing safety benefits (accident, injury, fatality) Terminal value of asset Average wages (% of wage for leisure) New research led to vast change in assumptions % of construction cost accepted method 1. Maintenance 1. 2. Construction costs 2. 3. Valuation Measurement 1. 2. 3. Misinterpretation of maint. quote (km v lane km) Clearer ex post (reduced strategic bias) • Disconcertingly large errors in estimation for a relatively straightforward project. ̵ Many potential sources of error ̵ Categorisation helps • Importance of generating independent values for costs (strategic bias minimisation) • Should one in ten ex ante CBAs be subject to in medias res or ex post follow-up and comparison to ex ante analysis? “Prediction is very difficult, especially about the future.” Niels Bohr (attributed) • Time spent thinking about possible sources of error is time well spent – framework available • At pre-project stage think not only about which costs and benefits but how they will be measured ..and ensure data stream exists − Programme design to enable measurement • Understanding what factors a CBA is sensitive to (sensitivity analysis) – presenting scenarios seems sensible!