Willingness to pay for BMP-induced water quality benefits and deviations around expected water quality outcomes

Figures & data

Table 1. Summary statistics of variables.

Variable^a	Mean	Standard deviation	Min.	Max.
Mean phosphorus reduction (PhoM)	20.058	15.689	0	40
Mean coliform reduction (ColM)	16.697	13.568	0	35
Standard deviation phosphorus reduction (PhoSD)	10.015	7.832	0	20
Standard deviation coliform reduction (ColSD)	6.654	5.759	0	15
Farmers’ cost^b	10.765	14.544	–5	35
Taxpayers’ cost^b	6.661	7.139	0	18
“Town people should pay” Likert scale measure (town)	8.632	2.484	2	12
Taxpayers’ cost × town	57.48	66.19	0	216
Age of farmer (age)	50.336	11.911	24	78

^a All six attribute characteristics and the interaction variables are divided by 10 to reduce the size of their coefficients in the econometric estimations.

^b Farmers’ cost and taxpayers’ cost are measured in Canadian dollars ($CAD).

Table 2. Mixed logit estimation of choice models with and without risk.

Variable	Model 1		Model 2
	(with risk attribute)		(no risk attribute)
	Random coefficients
Phosphorus’ reduction mean (PhoM)	0.7297***		0.4658***
	(0.2189)		(0.1735)
Coliform’s reduction mean (ColM)	0.9248***		0.5594***
	(0.3216)		(0.2158)
Phosphorus’ reduction deviation (PhoSD)	–0.2614		–
	(0.4390)
Coliform’s reduction deviation (ColSD)	–0.8486		–
	(0.5843)
Heterogeneity in mean coefficients
PhoM-age	–0.0082**		–0.0018
	(0.0041)		(0.0033)
ColM-age	–0.0123**		–0.0028
	(0.0062)		(0.0041)
PhoSD-age	0.0148*		–
	(0.0084)
ColSD-age	0.0231**		–
	(0.0111)
Non-random coefficients
Farmers’ cost	–0.4344***		–0.3417***
	(0.0445)		(0.0331)
Taxpayers’ cost	–1.7749***		–1.4279***
	(0.2544)		(0.2089)
Taxpayers’ cost × town	0.0763***		0.0574**
	(0.0279)		(0.0235)
Standard deviations of random coefficients
Phosphorus’ reduction mean (PhoM)	0.2798***		0.1677**
	(0.0896)		(0.0751)
Coliform’s reduction mean (ColM)	0.4040**		0.1550
	(0.1778)		(0.1781)
Phosphorus’ reduction deviation (PhoSD)	0.7215***		–
	(0.1931)
Coliform’s reduction deviation (ColSD)	0.7200***		–
	(0.2667)
Statistics
Log likelihood	–820.6073		–860.9241
LR chi^2(4)	465.9238		385.2901^a
Prob > chi^2	0.0000		0.0000
Num. obs.	959		959

^***

^**

^* = significance at 1, 5 and 10% levels, respectively.

^aThe degree of freedom for this model is 2.

Table 3. Covariances between mixed logit coefficients.

Estimated covariance matrix	PhoM	ColM	PhoSD	ColSD
Mean phosphorus reduction (PhoM)	0.078	–0.090	0.125	0.070
Mean coliform reduction (ColM)		0.163	–0.246	–0.236
Standard deviation phosphorus reduction (PhoSD)			0.521	0.364
Standard deviation coliform reduction (ColSD)				0.518
t-statistics for estimated covariances	PhoM	ColM	PhoSD	ColSD
Mean phosphorus reduction (PhoM)	1.67	1.92	1.60	0.76
Mean coliform reduction (ColM)		1.63	1.66	1.31
Standard deviation phosphorus reduction (PhoSD)			1.80	1.09
Standard deviation coliform reduction (ColSD)				1.82

The t-statistics were computed using the simulated method of Krinsky and Robb (1986).

Table 4. Willingness to pay (WTP) estimates for two age groups (5000 draws accounting for correlations between random coefficients).

	Younger farmers			Older farmers
	(= 30 years old)			(= 60 years old)
	Mean	CI^a, 10%		Mean	CI, 10%
Mean phosphorus reduction (PhoM)	1.1081	[0.061; 2.165]		0.5412	[–0.504; 1.598]
Mean coliform reduction (ColM)	1.2797	[–0.253; 2.802]		0.4290	[–1.104; 1.952]

^a CI = confidence intervals.

Figure 1. Willingness to pay (WTP) for phosphorus reduction.

Figure 2. Willingness to pay (WTP) for coliforms reduction.

Krinsky, I., and L. A. Robb. 1986. On approximating the statistical properties of elasticities. The Review of Economics and Statistics 68(4): 715–719.10.2307/1924536

 Web of Science ®Google Scholar