Moving Beyond Open Markets for Water Quality
Trading: The Gains from Structured Bilateral Trades
Tianli Zhao
Yukako Sado
Richard N. Boisvert
Gregory L. Poe
Cornell University
EAERE Preconference on Water Economics
Amsterdam, June 2009
Funding for this research was provided by the US-EPA Targeted Watershed Grant Program
(WS972841) and USDA Hatch Project funds provided through Cornell University (121-6415, 121-7810)
Neither the US EPA nor the USDA has reviewed or endorsed this publication, and the views expressed
herein may not reflect the views of these agencies.
Moving Beyond Open Markets for Water Quality
Trading: The Gains from Structured Bilateral Trades
Sado, Y. R.N. Boisvert, and G.L. Poe. Potential Cost Savings from Discharge
Allowance Trading: A Case Study and Implications for Water Quality Trading.
(http://www.eaere2009.org/)
Zhao, T., Y. Sado, R.N. Boisvert and G.L. Poe. “Open Markets” v. “Structured
Bilateral Trades”: Results of Economic Modeling of Point-to-Point Source
Water Quality Trading in the Non-Tidal Passaic River Basin
(http://aem.cornell.edu/profiles/poe.htm )
Outline:




Motivation, orientation and terminology,
Standard economic presentation of pollution allowance trading,
Adjusting standard economic presentation to reality of watershed,
Results.
Motivation: Water Quality Trading?

Water quality trading has had mixed success in the
United States.

Substantial financial and technological support by the US
EPA, and more than 48 established and pilot programs
• “Only 100 facilities have participated in trading” (US EPA 2008, E-S 1).
• Moreover, 80 of these 100 trades have occurred in a single
market, the Long Island Sound Nitrogen Credit Exchange
program.

Limited correspondence between open-market trading
system envisioned in standard economics
presentations and the institutional and hydrologic
circumstances in “typical” watersheds.
Orientation: The
Non-Tidal Passaic
River Basin
22 wastewater treatment plants
(WWTPs) required to reduce
total phosphorus concentrations
from present average of 2.33
mg/L to a target of 0.4 mg/L.
Trading Program Terminology

Open Markets: The standard theoretical presentation in
which firms buy and sell pollution allowances based on
marginal abatement costs through a price mechanism.


Best suited for:
• Large numbers of potential traders, with
• Heterogeneous abatement technologies across firms, and
• Heterogeneous present capacity to meet standards.
Structured Bilateral Trades: Profitable trading
opportunities identified, and implemented with multiyear
contracts between firms.

Best suited for:
• Small number of potential traders, with
• Homogeneous abatement technologies across firms, and
• Most, if not all, firms do not have present capacity to meet
standard.
Standard Economic Presentation (e.g., Tietenberg 2006)
n
Minimize
i 1
C i (e
0
i
ei )
Subject to:
n
i 1
d ij ei
Ej
( j 1, 2 ... n)
0
i
ei [0, e ]
eio = primary effluent level of source i,
ei = final effluent level of source i,
Ci (ei0 – ei) = abatement cost for source i.
Ej = aggregate emissions levels allowed at receptor j,
eij = amount of emissions at receptor j after discharger at source
i emits ei.
d ij
eij / ei
Standard Economic Presentation (e.g., Tietenberg 2006)
j
MCi (ei )
1 Ci (ei )
d ij
ei
1 Ck (ek )
d kj
ek
MCk (ekj )
Water Quality Trading (Hung and Shaw, JEEM, 2005)
n
Minimize
i
0
C
(
e
i
1 i
ei )
Subject to:
ei
ei
i 1
0
i
k 1
0
i
[0, e ]
n
d ki
ki
ki
,
ik
k 1
ik
0
Ti0 = zonal load cap taking into account background/natural levels of
pollutant and inflow from upstream sources adjusted for transfer
coefficient,
Tki = Tradable discharge allowances (TDAs) purchased by zone i from
zone k,
Tik = TDAs sold by zone i to zone k,
dki = Transfer coefficient (trading ratio) 0 < dki < 1
Two Critical Issues: Theory v. Real World
1. Ci (eio – ei)
a. Consistent with the previous marginal cost approach, “increasing
and strictly convex” (Hung and Shaw), “convex and twice
differentiable” (Montgomery), “continuous cost function”
(Tietenberg).
b. “Generally, pollution controls are feasible to implement in
relatively large installments that [can] reduce multiple units of
pollutants. Point sources in particular tend to purchase additional
loading reduction capability in large increments. For example a
wastewater treatment plant upgrade or plant expansion may be
designed to treat millions of gallons a day” (US EPA, 1996, p. 3-2).
2. The structure of dki
a. Downstream trades only.
b. Management Area approach to be applied to Upper Passaic
River Basin (Obrupta, Niazi and Kardos, JEM 2008).
1. Total Firm Costs: O&M
Data taken from 104 observations in “Nutrient Reduction
Technology Cost Estimations for Point Sources in the
Chesapeake Bay Watershed” (2002) and “Estimation of Costs of
Phosphorus Removal in Wastewater Treatment Facilities:
Adaptations of Existing Facilities” (2005).

For a given flow level, O&M cost rise with treatment level.
80

O&M Cost of Treatment 1
Additional Unit ($)

70
60
O&M treatment
costs fall with flow
level .
(e.g., 0.4 mg/L)
50
40
30
20
10
0
0
2
4
6
8
Flow (MGD)
10
12
14
1. Total Firm Costs: Capital Costs


Data taken from same sources as O&M data.
Discrete function with CC steps at 1 mg/L, 0.5 mg/L, 0.25 mg/L, 0.1
mg/L.
Capital Costs ($ Million)
30
25
20
~1.3 MGD (chemical treatment)
15
~13 MGD (chemical treatment)
10
5
0
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1.1
Concentration (mg/L)
CC rise with flow level (however large firms will have lower average capital costs).
2. Management
Areas
Pompton
Lower
Passaic
Upper
Passaic
22 wastewater treatment plants
(WWTPs) required to reduce
total phosphorus concentrations
from present average of 2.33
mg/L to a target of 0.4 mg/L.
2: Structure of Management Areas/Endpoints
2a.
Downstream trades only, each firm is own management area.
Downstream trades only (seller located upstream),
Water quality standard met at all points in watershed, no “hot spots”,
Firms with lowest abatement costs must be located upstream.



2b.
Three management areas two endpoints.

Non-Tidal Passaic River Basin TMDL, 2008.
Buyer
Upper Passaic
MA
Pompton
MA
Lower Passaic
MA
Upper Passaic MA
Yes
No
Yes
Pompton MA
Yes
Yes
Yes
Lower Passaic MA
No
No
Yes
Seller
Note: upstream trades between management areas not allowed.
2a: Trading Ratios (Downstream Trades Only)
Trading Ratios: Seller always removes one more lb.

1, buyer receives 1 lb. allowance.
0.780, buyer receives 0.780 lbs. allowance.
0 (or empty cell), no direct physical linkage or “buyer” located
upstream.



TRADING RATIOS
Sellers in rows, buyers in columns
D1
D1
D2
D3
P1
P2
P3
P4
P5
P6
P7
P8
W1
W2
W3
R1
W4
WQ
T1
T2
P9
P10
P11
D2
1
D3
1
1
1
1
1
P1
0.854
0.854
0.854
1
P2
0.854
0.854
0.854
1
1
P3
0.854
0.854
0.854
1
1
1
P4
0.854
0.854
0.854
1
1
1
1
P5
0.780
0.780
0.780
0.780
0.780
0.780
0.780
1
P6
0.780
0.780
0.780
0.780
0.780
0.780
0.780
1
1
P7
0.780
0.780
0.780
0.780
0.780
0.780
0.780
1
1
1
P8
0.780
0.780
0.780
0.780
0.780
0.780
0.780
1
1
1
1
W1
W2
1
W3
1
1
R1
W4
1
1
1
WQ
T1
T2
P9
P10
P11
1
1
1
1
1
1
1
1
0.720
0.720
1
1
1
1
1
2b: Trading Ratios (Three Management Areas)

Trading Ratios: Seller always removes one more lb.
1, buyer receives 1 lb. allowance.


0.780, buyer receives 0.780 lbs. allowance.
1.093, buyer receives 1.093 lbs. allowance.

0 (or empty cell), trades not allowed.

TRADING RATIOS
Sellers in rows, buyers in columns
D1
D2
D3
P1
P2
P3
P4
P5
P6
P7
P8
W1
W2
W3
R1
W4
WQ
T1
T2
P9
P10
P11
D1
D2
D3
1
1
1
0.900
0.900
0.900
0.900
0.923
0.923
0.923
0.923
0.821
0.821
0.821
0.649
0.821
0.486
0.486
0.608
1
1
1
0.900
0.900
0.900
0.900
0.923
0.923
0.923
0.923
0.821
0.821
0.821
0.649
0.821
0.486
0.486
0.608
1
1
1
0.900
0.900
0.900
0.900
0.923
0.923
0.923
0.923
0.821
0.821
0.821
0.649
0.821
0.486
0.486
0.608
P1
0.854
0.854
0.854
1
1
1
1
0.900
0.900
0.900
0.900
0.797
0.797
0.797
0.630
0.797
0.462
0.462
0.577
P2
0.854
0.854
0.854
1
1
1
1
0.900
0.900
0.900
0.900
0.797
0.797
0.797
0.630
0.797
0.462
0.462
0.577
P3
0.854
0.854
0.854
1
1
1
1
0.900
0.900
0.900
0.900
0.797
0.797
0.797
0.630
0.797
0.462
0.462
0.577
P4
0.854
0.854
0.854
1
1
1
1
0.900
0.900
0.900
0.900
0.797
0.797
0.797
0.630
0.797
0.462
0.462
0.577
P5
0.780
0.780
0.780
0.780
0.780
0.780
0.780
1
1
1
1
0.771
0.771
0.771
0.600
0.771
0.429
0.429
0.536
P6
0.780
0.780
0.780
0.780
0.780
0.780
0.780
1
1
1
1
0.771
0.771
0.771
0.600
0.771
0.429
0.429
0.536
P7
0.780
0.780
0.780
0.780
0.780
0.780
0.780
1
1
1
1
0.771
0.771
0.771
0.600
0.771
0.429
0.429
0.536
P8
0.780
0.780
0.780
0.780
0.780
0.780
0.780
1
1
1
1
0.771
0.771
0.771
0.600
0.771
0.429
0.429
0.536
W1
0.872
0.872
0.872
0.900
0.900
0.900
0.900
0.956
0.956
0.956
0.956
1
1
1
0.692
1.000
0.500
0.500
0.625
W2
0.872
0.872
0.872
0.900
0.900
0.900
0.900
0.956
0.956
0.956
0.956
1
1
1
0.692
1.000
0.500
0.500
0.625
W3
0.872
0.872
0.872
0.900
0.900
0.900
0.900
0.956
0.956
0.956
0.956
1
1
1
0.692
1.000
0.500
0.500
0.625
Note: 90% Worst case trading ratio used in TMDL
R1
W4
0.996
0.872
0.996
0.872
0.996
0.872
1.029
0.900
1.029
0.900
1.029
0.900
1.029
0.900
1.093
0.956
1.093
0.956
1.093
0.956
1.093
0.956
1.029
1
1.029
1
1.029
1
1 0.69231
1.029
1
0.621
0.500
0.621
0.500
0.755
0.625
WQ
1
1
0.829
T1
T2
1
1
0.829
0.720
0.720
1
P9
0.537
0.537
0.537
0.566
0.566
0.566
0.566
0.610
0.610
0.610
0.610
0.523
0.523
0.523
0.421
0.523
0.290
0.290
0.363
1
0.871
0.871
P10
0.396
0.396
0.396
0.418
0.418
0.418
0.418
0.450
0.450
0.450
0.450
0.386
0.386
0.386
0.311
0.386
0.214
0.214
0.268
0.664
1
1
P11
0.396
0.396
0.396
0.418
0.418
0.418
0.418
0.450
0.450
0.450
0.450
0.386
0.386
0.386
0.311
0.386
0.214
0.214
0.268
0.664
1
1
Baseline, No Trade Scenario

All firms must operate at 0.4 mg/L phosphorus
corresponding to TMDL and their historic average flow
levels.


Flow and concentration levels converted to pounds.
Annualized total costs:
O&M
CC
Total

Total capital investment -
$1,924,000
$1,234,000
$3,159,000
$12,811,000
Percent Total Cost Savings
Open
Market


Downstream trades only, each firm is
own management area.
1.19 %
Three management areas, two
endpoints
2.64 %
Structured
Bilateral
Trades
Overall gains from trade low in open market.
Gains increase (marginally) with greater flexibility in trading (i.e. between
rows in table above)
Percent Total Cost Savings


Open
Market
Structured
Bilateral
Trades
Downstream trades only, each firm is
own management area.
1.19 %
5.82 %
Three management areas, two
endpoints
2.64 %
16.09 %
Overall gains from trade low in open market.
Gains increase marginally with greater flexibility in trading (i.e. between
rows in table above)
 Substantial gains of trade relative to open market scenario captured by not
having many firms upgrade beyond 0.5 mg/L. (i.e. between columns in table
above)
 Savings in bottom right cell attributed to upgrades of critically located, large
waste treatment plants, who sell allowances to smaller firms so as to
preempt upgrading.
Why Structured Bilateral Trades Work

Bear Creek, CO, each year a large discharger
(Evergreen Metro) reduces phosphorus release in a
trade of 40-80 pounds per year so that a smaller
discharger (Forest Hills) does not have to undergo a
costly upgrade to its facilities.
“It is estimated that Forest Hills saves over $1.2 million,
the cost of an expensive system replacement that would
be necessary to meet their allocation without a trade… In
exchange for Evergreen Metro reducing their discharge,
Forest Hills pays an undisclosed amount of money that
has been estimated to be around $5,000 per year”
- Breetz et al., p. 28
Example of Sequential
Restrictions
WWTP
Restriction % Total Cost
Saving
Open-Market Trade
2.64
W1
> 0.50 mg/L
3.95
P7
> 0.50 mg/L
4.60
P1
> 0.50 mg/L
5.44
P6
> 0.50 mg/L
6.24
W3
> 0.50 mg/L
7.41
D1
> 0.50 mg/L
7.95
W2
> 0.50 mg/L
9.68
P3
> 0.50 mg/L
10.61
P5
> 0.50 mg/L
11.48
P2
> 0.50 mg/L
11.99
R1
> 0.50 mg/L
13.26
P4
> 0.50 mg/L
13.55
D2
> 0.50 mg/L
13.88
D3
> 0.50 mg/L
14.32
P11
> 0.50 mg/L
15.23
P10
> 0.50 mg/L
16.09
In This Example: 19 “Structured Bilateral Trades”
Seller
Buyer
W4
W1, W2, W3
R1
P1, P2, P3, P4, P6
P8
D1,
P5, P7
T2
D1, D2, D3
P9
P10, P11
WQ
R1
T1
R1
Green type indicates sellers that will have to upgrade.
Summary and Conclusions

Open market trading program based strictly on independent
firms making purchase/sales decisions by comparing market
price with marginal abatement cost is:



Unlikely to develop in this watershed given the limited number of
WWTPs and other characteristics.
Likely to provide only relatively modest cost savings in this
watershed if an open-market trading program develops.
A structured bilateral trading program in which operating and
capital costs are explicitly considered and a limited number of
multi-year contracts established between firms has the
potential to generate larger watershed-wide gains from trade
and cost savings in this watershed.
Extra Slides Below for Reference
Not Part of Presentation
Moving Beyond Open Markets for Water Quality
Trading: The Gains from
Structured Bilateral Trades in Small Watersheds
Sado, Y. R.N. Boisvert, and G.L. Poe, Potential Cost Savings from Discharge
Allowance Trading: A Case Study and Implications for Water Quality Trading.
(http://www.eaere2009.org/)
Zhao, T., Y. Sado, R.N. Boisvert and G.L. Poe. “Open Markets” v. “Structured
Bilateral Trades”: Results of Economic Modeling of Point-to-Point Source
Water Quality Trading in the Non-Tidal Passaic River Basin
(http://aem.cornell.edu/profiles/poe.htm )
Discontinuous Fixed Costs
Full Cost Economic Optimum