Actual averting expenditure versus stated willingness to pay
The purpose of this study is to perform a complete comparison of actual averting expenditure and stated willingness to pay measures, and to determine if the averting expenditure is a lower bound of the willingness to pay measured from contingent valuation experiment as suggested by literature. In addition to the single value comparison, Bootstrap, Krinsky and Robb, Jackknife, and Cameron are four simulation methods used to calculate confidence intervals for response function. Sample sizes of 100, 200, and 1000 are simulated 100 and 200 times respectively. A set of data with 540 households from a contingent policy referendum survey is employed for our purpose. Under a specific level of BOD improvement, a one-to-one single mean value comparison of the actual averting expenditure is greater than the mean willingness to pay from utility difference model. The empirical results are consistent with the theoretical expectation for expenditure difference that averting expenditure is a lower bound of willingness to pay generated from the contingent valuation method. A confidence interval, which contains the true mean willingness to pay at least 90% of the times, includes the actual averting expenditure as a lower bound of the mean willingness to pay as theory predicts.