Ferraro, P.J. Forthcomng. Integratng Bophyscal and Economc Informaton to Gude Land Conservaton Investments. In Economc Advances n Land Use Polcy: Implcatons of Emergng Research for Conservaton of Farm, Forest, and Rural Character, edted by Robert Johnson and Stephen Swallow. Resources for the Future Press, Washngton, D.C. Integratng Bophyscal and Economc Informaton to Gude Land Conservaton Investments Paul J. Ferraro Department of Economcs Andrew Young School of Polcy Studes Georga State Unversty P.O. Box 3992 Atlanta, GA 30302-3992 (Voce) 404-651-1372; (Fax) 404-651-0425 pferraro@gsu.edu * The author acknowledges helpful comments from two anonymous revewers and partcpants n the 2003 NAREA Land Use Polcy workshop. The author also thanks Lee Macbeth and the Department of Water at the Cty of Syracuse for access to the Lake Skaneateles data and Resources for the Future and the New York Water Insttute for fnancal support.
I. Introducton Concerns over the effect of prvate land use on the supply of envronmental amentes have led to an ncreasng global relance on conservaton contractng ntatves (Ferraro and Kss 2002). The term conservaton contractng descrbes the contractual transfer of payments from one party (e.g., government) to another (e.g., landowner) n exchange for land use practces that contrbute to the supply of an envronmental amenty (e.g., bodversty, water qualty). Examples of conservaton contracts nclude easements and short-term conservaton leases. A key ssue n the desgn of conservaton contractng ntatves, lke any conservaton polcy, s how to ntegrate nformaton about spatally varable bophyscal and economc condtons nto a cost-effectve conservaton plan. Much of the prevous work on targetng scarce conservaton funds has focused on the conservaton of bologcal dversty. Targetng approaches favored by bologcal scentsts and conservatonsts emphasze the envronmental amentes that a gven land unt produces, whle often gnorng the costs of acqurng those amentes. For example, Dobson et al. (1997), based on ther fndng that endangered speces n the Unted States were concentrated spatally, suggest conservatonsts should focus ther efforts on a small number of geographc areas. In response, Ando et al. (1998) assert that varablty n economc factors s just as mportant as ecologcal varablty n effcent speces conservaton, specfcally notng an approach that consders both economc and ecologcal varablty could cost less than one-sxth the cost of an approach consderng only ecologcal varablty. A smlar debate has developed over targetng ecosystem conservaton nvestments at the global scale (Mttermer et al. 1998; Balmford et al. 2000). Other studes by economsts have also demonstrated the mportance of ntegratng bophyscal 1
and economc data, as llustrated by Polasky et al. (2001) for the case of speces conservaton n Oregon, and Babcock et al. (1996, 1997) for the case of the Conservaton Reserve Program. Ths chapter extends these prevous analyses n several ways. Frst, the analyss focuses on an ncreasngly common, but lttle studed, conservaton ntatve: conservaton contractng for water qualty objectves. The results of the emprcal analyss support prevous emprcal work suggestng the falure to ncorporate cost data n conservaton nvestment decsons can lead to large effcency losses. Moreover, studes of cost-effcent targetng (e.g., Ando et al.; Polasky et al.; Babcock et al.) have tended to focus on a sngle bophyscal attrbute (e.g., speces absence or presence, erodblty of sol, dstance to water). A narrow focus on a sngle attrbute, however, fals to consder the full range of bophyscal attrbutes that are crtcal to the supply of an envronmental amenty. Most conservaton ntatves, lke the U.S. Conservaton Reserve Program (USDA 1999) or World Wldlfe Fund s Global 200 ntatve (Olson et al. 2000), dentfy multple bophyscal attrbutes of nterest. In the context of habtat protecton, Prendergast et al. (1999) pont out that practtoners and polcymakers rarely use the tools and results publshed n the academc lterature. In large part, these tools and results have not been adopted because ther development and applcaton do not take nto account the objectves and approaches of practtoners and polcymakers. To address ths oversght, the emprcal applcaton of ths chapter uses data avalable to decsonmakers, and consders explctly the actual approaches used by decson-makers n the feld. The problem s also approached at the geographc scale at whch decsons are beng made: ndvdual parcels rather than large admnstratve dstrcts or GIS polygons on the landscape. Unlke prevous work, we recognze that there s often lttle agreement about the approprate way to estmate the envronmental benefts provded by a sngle parcel and thus use 2
multple methods to gude the emprcal analyss. Fnally, there s ncreasng scentfc nformaton that suggests bophyscal thresholds are mportant when desgnng conservaton ntatves (e.g., a rparan buffer has lttle effect on water qualty unless t acheves a mnmum sze). Few analyses of conservaton nvestments, however, have ncorporated such thresholds (notable exceptons nclude Farzn 1996; Wu et al. 2000; and Bulte and van Kooten 2001). Ths chapter demonstrates how such thresholds can be ncorporated nto the decson-makng process. In the emprcal analyss, we compare the conservaton contract portfolos selected wth and wthout threshold constrants. In the next secton, the case study for the emprcal analyss s ntroduced. In secton III, the data are characterzed and the decson model s developed. The results of the emprcal analyss are then presented n secton IV. In secton V, the model of Secton III s adapted to ncorporate thresholds, and n Secton VI, the effects that thresholds have on the selecton of the optmal conservaton contract portfolo are examned. Secton VII explores reasons why economc approaches to targetng land conservaton nvestments are not appled n practce. II. Case Study: the Lake Skaneateles Watershed Program The use of conservaton contracts to acheve water qualty objectves s becomng an ncreasngly popular polcy tool (Johnson et al. 2001). For example, the New York Cty Watershed Management Plan wll spend $250 mllon on conservaton contractng wth prvate landowners n the Catskll-Delaware watershed over the next 10 years to protect the Cty s water supply and mantan ts fltraton waver from the Envronmental Protecton Agency (NRC 2000, 213-239). Examples of other contractng ntatves for water qualty nclude North Carolna s $30 mllon Clean Water Management Trust Fund, Massachusetts s $80 mllon effort to acqure 3
rparan land to protect Boston s Wachusett Reservor, and Costa Rca s $16 mllon per year effort to secure conservaton contracts n (among other areas) the watersheds of muncpal water supples and hydroelectrc dams. In partcular, scentsts and polcymakers have dentfed the establshment of vegetated rparan zones that protect surface waters from nputs of nutrents, pestcdes, eroded sol, and pathogens as an mportant polcy for mprovng water qualty (Tlman et al. 2001). One such rparan buffer acquston ntatve s currently underway n upstate New York. The Cty of Syracuse (populaton 163,860) obtans ts drnkng water from Lake Skaneateles, whch s located outsde of the Cty s regulatory jursdcton. The lake s 16 mles long, less than onemle wde on average, and has a 60 square mle watershed coverng three countes, seven townshps and one vllage. The populaton of the watershed s about 5,000 resdents, concentrated largely n the northern half of the lake where the Cty s ntakes ppes are located. Land use s manly a mx of forest (40%) and agrcultural land (48%), on whch croppng and dary farmng are most common. The water from the lake s of exceptonally hgh qualty and the Cty, usng only dsnfecton by chlornaton, meets drnkng water standards wthout coagulaton or fltraton. 1 In recent years, however, the Cty has come under ncreasng pressure to consder fltraton n order to satsfy the provsons of the Envronmental Protecton Agency s (EPA) Surface Water Treatment Rule. In 1994, the Cty sgned a Memorandum of Agreement (MOA) wth the New York State Department of Health allowng the Cty to avod flterng water from the lake. The MOA requres that the Cty commt to a long-term watershed management program to reduce pathogen, chemcal, nutrent and sedment loadng nto the lake. An mportant part of the 1 An estmated 20-65 mllon Amercans drnk unfltered surface water (DeZyane, 1990), ncludng ctzens n the ctes of New York, Boston and San Francsco. 4
management program s a conservaton easement acquston program through whch up to $5 mllon wll be spent over a seven-year perod (2001-2008) to secure easements on prvately owned rparan parcels. By securng easements on rparan buffers n the watershed, the Cty hopes to avod, or delay, the estmated $60-$70 mllon cost of a new fltraton plant. The Cty wants to allocate ts lmted budget across the watershed n a way that wll have the greatest effect on mantanng and mprovng water qualty n the lake (Myers et al. 1998). The focus of ths analyss s on prortzng the acquston of easements from an avalable populaton of 202 rparan parcels n the upper watershed of Lake Skaneateles. Bophyscal and economc data on these parcels were obtaned from the Geographc Informaton Systems database of the Cty of Syracuse s Department of Water. 2 The southwestern end of the lake s protected publc land and s thus excluded from the analyss. Data on parcels n the southeastern end of the lake were not avalable at the tme of analyss, but because these parcels are far from the Cty s ntake ppes, excludng them wll have only mnor effects on the fnal results. III. Case Study: data and conceptual approach Each rparan parcel n the watershed, when protected by an easement, s assumed to generate envronmental benefts, e, to the Cty of Syracuse at a cost of c + t, where c represents the reservaton prce of the landowner for acceptng an easement on hs or her property and t s the transacton cost assocated wth creatng and montorng a contract. The unt of analyss s the parcel. Wthn each parcel, envronmental benefts and costs are assumed unformly dstrbuted. In other words, each acre n the parcel s equally as valuable, whether measured for 2 These data, strpped of owner nformaton, can be downloaded at http://epp.gsu.edu/pferraro/research/ workngpaper/workngpapers.htm. 5
envronmental benefts or for productve uses. These are the same assumptons used by the Cty of Syracuse n ts easement acquston program. Beneft Data The Cty wshes to reduce sedment, chemcal, pathogen and nutrent loadng nto ts water supply. Sophstcated hydrologcal models, however, are not avalable for the Lake Skaneateles watershed. To measure the contrbuton of each parcel to the Cty s water qualty objectves, the Cty s Department of Water convened a scentfc panel to help t develop a parcel-scorng system based on known land attrbutes n the watershed (Myers et al. 1998). The panel developed two potental systems: an nterval-scale scorng equaton and a rato-scale scorng equaton. The equatons, whch are descrbed n the appendx, assgn a score to each parcel; the hgher the score, the hgher the beneft from easement acquston. Two other common parcel-scorng methods, the categorcal scorng system (smlar to that used by the U.S. Conservaton Reserve Program) and the Parcel-Pollutant-Weghtng (PPW) model (Azzano et al. 2002), are also used n the emprcal analyss and are descrbed n the appendx. The nterval-scale, rato-scale, and categorcal scorng equatons use the same bophyscal characterstcs but wegh and normalze them dfferently (the characterstcs have to be normalzed so that parcel scores are not fundamentally altered by changes n the characterstcs unts of measurements). The Parcel-Pollutng-Weghtng scorng equaton dffers from the other three because t combnes nformaton on bophyscal characterstcs and results from polluton modelng n order to score each parcel. Nether theory nor extant emprcal evdence argue for the superorty of one scorng method over the others. 6
All four beneft-measurng methods generate parcel scores ether from weghted lnear functons of the attrbutes or by assgnment of ponts to each parcel based on ts bophyscal attrbutes or land uses. Such scorng methods are qute common n the academc lterature (e.g., Voogd 1983; Lemunyon and Glbert 1993), n federal agency gudelnes (e.g., USFWS 1981; Terrell et al. 1982; Allen 1983; McMahon 1983; Allen and Hoffman 1984), n water qualty protecton ntatves (e.g., Smth et al. 1995; Rowles and Stlnger 1999; MDC 1999; Hruby et al. 2000; FDEP 2000), and n the mult-bllon dollar conservaton efforts of the U.S. Conservaton Reserve Program (Feather et al. 1998), land trusts (e.g., The Nature Conservancy; Master 1991), nternatonal habtat protecton groups (e.g., World Wldlfe Fund; see Olson et al.), natonal wldlfe protecton ntatves (e.g., Partners n Flght, documented by Carter et al. 1999), and farmland protecton ntatves (e.g., Amercan Farmland Trust). 3 In the absence of sophstcated hydrologcal models for the Skaneateles watershed, t s not possble to determne whch of the four parcel scorng methods s best. 4 If there s postve correlaton among the dfferent scorng methods (whch would be expected f they are all attemptng to measure the same amenty), a smple approach to prortzng easement acquston would be to dentfy the optmal buffer portfolos selected under several scorng methods and then dentfy a set of hgh-prorty parcels that ncludes only parcels found n every portfolo (.e., parcels n common wthn each optmal portfolo across the parcel-scorng methods). Ths approach s appled n secton IV. As observed from Table 1, the Spearman correlatons among the parcel scores assgned by each scorng method are strongly postve. 3 Ferraro (2004) explores an alternatve way to assgn prorty to conservaton nvestments when the ecologcal benefts from nvestment cannot be collapsed nto a sngle value. 4 Even f sophstcated models exsted for estmatng sedment, chemcal, pathogen and nutrent loadng, one would have to somehow combne these measures to derve a measure of water qualty benefts from an easement on a gven parcel. 7
[Table 1 about here] Cost Data There were not enough observatons on sales of propertes wth easements n the regon to estmate a hedonc equaton of easement costs. A regonal apprasng company (Gardner 2000) estmated that the Cty of Syracuse would have to pay between 40% and 60% of a parcel s assessed land value to obtan an easement. An estmate of 50% s used n ths analyss. A change n the percentage would affect only the number of parcels that can be acqured for a gven budget, not the order n whch the parcels are acqured. Based on transacton cost nformaton from the Fnger Lakes Land Trust, whch operates n the regon, a transacton cost of $5000/easement s also assumed. Varyng the transacton cost from $2500 to $12,500 dd not generate dramatc changes n the parcel rankngs. 5 Future analyses can ncorporate new nformaton on costs gathered by practtoners n the course of contactng landowners. The Cty of Syracuse may also want to consder usng a procurement aucton to solct reservaton prces from landowners (Cummngs et al. 2003). Optmal Easement Portfolo Selecton Problem The Cty of Syracuse s optmal easement acquston program can be vewed as maxmzng the total beneft score subject to a budget constrant (see appendx for a more formal representaton). Ths maxmzaton problem s equvalent to rankng parcels from hghest to lowest based on ther e /(c +t ) rato and acceptng contracts untl the budget s exhausted. 5 The exceptons were a few small, nexpensve parcels for whch a change n transacton costs can have a large relatve effect on easement cost. 8
The Cty of Syracuse, however, dd not formulate ts approach to easement acquston n ths manner. Lke many conservaton ntatves (e.g., Mttermer et al.), the Cty planned to allocate ts funds by rankng parcels from the hghest score (e ) to the lowest and acqurng easements untl the budget was exhausted. In ths approach, there s a crtcal level of envronmental beneft, e, for whch all parcels wth more formal representaton). e > e are contracted (see appendx for a The Cty s prortzaton formulaton gnores the opportunty costs of contracted parcels and, as suggested by prevous emprcal analyses (refer to ctatons n the ntroducton), ts portfolo for any gven budget wll generate lower beneft scores than the portfolo generated from the optmal portfolo. How much lower s an emprcal queston. IV. Case Study: emprcal results The Cty plans to spend $1 - $2.5 mllon dollars and then evaluate whether further easement acqustons are requred. We therefore solve the optmal easement portfolo problem under each scorng method for budgets of D = $1 mllon and D = $2.5 mllon (maps of the correspondng optmal portfolos can be found n Ferraro 2002). Table 2 presents, for each beneft-scorng method, the percentage of total envronmental benefts avalable n the watershed that are secured by the optmal portfolo and the percentage of total envronmental benefts avalable n the watershed that are secured by the portfolo that gnores opportunty costs (.e., funds are allocated based on beneft scores alone). Consstent wth prevous research, large effcency losses are assocated wth gnorng costs n the fundng allocaton decson. For a budget of $1 mllon, the beneft-only approach acheves 16 to 42 percent of what the optmal approach acheves; for a budget of $2.5 mllon, t acheves 36 to 65 9
percent of what the optmal approach acheves. The large effcency gans from ncorporatng economc costs explctly n the decson-makng derve from the moderate postve correlaton between beneft (e ) and cost (c ) measures and the greater relatve heterogenety of costs compared wth that of benefts (Ferraro 2003; more on ths at the end of ths secton). Whle the formulaton that ntegrates beneft and cost data s clearly benefcal, each scorng method generates a dfferent optmal portfolo. As mentoned n the prevous secton, one way to proceed would be to dentfy the parcels that are selected for acquston under all four scorng methods. These parcels mght be regarded as "hgh prorty" for an easement acquston program because they are found n all four optmal buffers. Such an approach would ft well wth the Cty of Syracuse s approach to easement acquston. Although the Cty has estmated that t mght spend up to $5 mllon for easement acquston, t plans to begn acqurng easements sequentally and perodcally evaluate whether or not more easements wll need to be acqured. Thus the Cty wants to know wth whch parcels t should begn ts acquston efforts. The set of hgh prorty parcels would be a reasonable place to start. For any gven avalable budget, one can dentfy a set of prorty parcels that exhausts the budget by changng the value of the budget under whch the optmal buffers are derved. [Table 2 about here] For example, solvng for the portfolos when the budget s $1 mllon, 11 parcels are found n each of the four optmal buffer solutons and these easements can be acqured for $210,900. Solvng for the portfolos when budget s $2.5 mllon, 46 parcels are found n each of the four optmal buffer solutons and these easements can be acqured for $1,445,150. Table 3 demonstrates how well the hgh prorty set of parcels performs compared to the optmal portfolos chosen under the four scorng equatons when the budget s $210,900 and $1,445,150. 10
For example, the hgh-prorty portfolo, were ts parcels to be scored accordng to the ntervalscale scorng equaton, acheves 92% of the benefts that are acheved by the optmal portfolo at a budget of $1,445,150. The data n Table 3 suggest that even f one of the scorng equatons were the true measure of parcel benefts, the Cty of Syracuse would not lose a substantal amount of effcency by selectng the hgh-prorty portfolo of parcels. [Table 3 about here] Thus, as n prevous analyses that looked at conservaton nvestments n other contexts, we fnd that ntegratng both cost and beneft nformaton explctly nto the decson-makng n Lake Skaneateles can be vtal to ensurng that scarce funds go as far as they can toward achevng polcy objectves. The costs of acqurng and analyzng such nformaton, however, can be substantal. Under whch condtons s ntegratng cost and beneft nformaton lkely to be vtal to effectve decsonmakng, and under whch the condtons wll the falure to use both cost and beneft data result n lttle, f any, losses n effcency? Assgnng prorty to land conservaton nvestments on the bass of bophyscal data alone would be approprate only f (1) benefts and costs were negatvely correlated across stes and (2) the relatve spatal varablty of benefts was greater than the relatve spatal varablty of costs. If these two condtons do not hold, as they do not n Lake Skaneateles, an approach that gnores the heterogenety of conservaton costs across stes would perform poorly n ensurng that every dollar spent acheves the maxmum envronmental benefts that were possble. Thus, n a habtat restoraton program, for example, we would expect that the greater the postve spatal correlaton between envronmental benefts and restoraton costs, and the greater the spatal varablty of restoraton costs compared wth the varablty of envronmental benefts, the greater wll be the effcency losses f conservaton agents gnore costs when 11
makng decsons on where to restore habtat. Even f costs and benefts were negatvely correlated, but relatve cost varablty was much greater than relatve beneft varablty, there could be large gans from ntegratng restoraton costs nto the prortzaton process. In the case of nvestments n endangered speces recovery n the Unted States, Ferraro (2003) demonstrates that the polcy context has attrbutes smlar to the Cty of Syracuse case: (1) benefts, reflected n the prorty scores, and recovery costs are postvely correlated, and (2) recovery costs are more varable than prorty scores across speces. In such a stuaton, allocatng funds based on prorty scores alone would be qute neffcent. In fact, Ferraro shows that polcy analysts examnng expendture decsons n ths envronment may fnd lttle or no postve relatonshp between beneft measures and the extent and lkelhood of fundng for speces recovery when funds are beng spent to maxmze the envronmental benefts of every dollar spent. Data from Balmford et al. (2003) suggest that, at the global scale, the costs of habtat protecton are spatally negatvely correlated wth the spatal benefts (usng brd speces densty as a proxy for the benefts of habtat protecton). Hgh bodversty areas are typcally found n low-ncome natons n whch the opportunty costs of conservaton are low. In such a context, focusng only on benefts may not yeld a substantally neffcent conservaton nvestment portfolo. The authors, however, also note that the habtat acquston cost measures are much more varable than the beneft measures. Thus, ther data suggest that gnorng beneft data entrely and acqurng more detaled cost data wth whch to target global habtat conservaton nvestments may be the most effectve way of spendng scarce conservaton funds. The same deas can be appled to contexts n whch benefts are gnored and only economc data are used to target conservaton funds. For example, Ferraro (2003) examnes 12
Georga s Envronmental Protecton Dvson (EPD) 2001 rrgaton aucton, n whch the state compensated farmers who voluntarly agreed to stop rrgatng ther crops durng the year. The budget was not suffcent to pay all farmers n the regon and thus the EPD asked economsts at Georga State Unversty to desgn an aucton to allocate the state s scarce procurement budget (Cummngs et al. 2003). Gven that Georga s rrgaton water s not metered, and tme and money for data collecton were lmted, a decson was made to allocate the no-rrgaton contracts accordng to cost alone. Farmers bd the amount of money per acre they were wllng to accept to forgo rrgaton on ther lands, the bds were ordered from lowest to hghest and the state procured contracts untl the budget was spent. The decson to focus only on cost measures was justfed based on agronomc expertse that suggested water use and contract costs are negatvely correlated and thus the parcels that experence the greatest water use are also the low-cost parcels. Even wth postve correlaton between water use and contract costs, however, a hgh relatve varablty of water use compared to the relatve varablty of contract costs could have greatly decreased the cost-effcency of the aucton that assgned prorty to contracts on the bass of cost alone (no data on the spatal varablty of antcpated water use among aucton partcpants exst). V. Thresholds: concepts and problem formulaton The emphass on parcel-level attrbutes n the analyss above may be napproprate f there exst thresholds of rparan buffer area below whch lttle, f any, water qualty protecton can be expected. The mportance of bophyscal thresholds n conservaton polcy desgn has been noted n a varety of contexts, ncludng endangered speces conservaton (Shaffer 1981; 13
Lande 1987; Wu et al. 2000) and water qualty protecton (Schueler 1994, 1995; Zoner and Lmtz 1994; Wang et al. 1997, 2000), but only a few economc analyses have ncorporated bophyscal thresholds (e.g., Farzn 1996; Wu et al. 2000; Bulte and van Kooten 2001). Ignorng threshold effects, partcularly when the avalable budget s small, may result n a substantal loss of envronmental benefts. Interventons wll be scattered over the landscape and fundng levels n any gven target area may be nadequate to reach the threshold needed to mantan current water qualty levels or to acheve sgnfcant envronmental mprovements. In an emprcal study, Wang et al. (1997) found (1) ndcators of water qualty were negatvely correlated wth the amount of agrcultural land n the entre watershed and n a 100- meter-wde buffer along streams; 6 and (2) the relatonshp between agrcultural land and water qualty was nonlnear a substantal declne n water qualty occurred after agrcultural land use exceeded 50%. Wth more ntensve agrcultural use or urban uses, the threshold value decreased to between 10% and 20%. A recent EPA (1999) report noted that thresholds for a declne n water qualty can take the form of sze and amount of rparan buffer zones. Condton of rparan zones and changes n percent of buffer areas can ndcate a declne n water qualty due to sol eroson, sedment loadng, and contamnant runoff. However, there have been no general rules of thumb developed specfcally for rparan areas. Consequently, the emprcal analyss below s ntended to demonstrate a way n whch bophyscal thresholds can be ncorporated nto the targetng process, rather than to clam such thresholds exst n the Lake Skaneateles watershed. The Lake Skaneateles upper watershed s made up of 16 sub-watersheds, or catchments. The Cty has determned that each easement wll be desgned to secure a 100-foot-wde rparan buffer along the entre stream length of the property. The next secton examnes the effect of 6 Correlatons were generally stronger, however, for the entre watershed than for the buffer. 14
mposng a threshold requrement on the area of 100-foot-wde rparan buffer n a gven catchment. Emprcally, the threshold s examned at three levels: 50%, 80% and 90% of the avalable rparan buffer n the catchment. For example, f there s a 50% threshold, no water qualty benefts can be acheved n a catchment through conservaton contractng unless at least 50% of the avalable 100-foot-wde rparan buffer s protected through easements. Thus a decson-maker must now select not only the parcels on whch to establsh a conservaton contract, but also the catchments n whch to establsh contracts. Unlke the targetng approach n Secton III, targetng conservaton nvestments costeffcently n the presence of threshold effects s not as smple as rankng parcels by ther beneftcost rato. One must have some basc understandng of constraned optmzaton and lnear programmng. However, the approach used below and outlned n the appendx can be mplemented n Mcrosoft Excel s Solver algorthm, whch allows practtoners a relatvely straghtforward way to program a constraned optmzaton problem. VI. Thresholds: emprcal results As n secton IV, we solve for the optmal easement portfolo under each scorng method for budgets of D = $1 mllon and D = $2.5 mllon. As one would expect, threshold constrants result n spatal concentraton of contracts on the landscape (spatal representaton of the solutons can be found n Ferraro 2002). Table 4 presents the percentage of parcels n the optmal buffer portfolo that ncorporates thresholds that were also found n the optmal portfolo derved wthout threshold constrants. [Table 4 about here] 15
For a gven scorng method, the spatal concentraton effect of thresholds on the optmal contract portfolo s generally greatest at low budget levels and hgh thresholds. For example, usng the PPW scorng method wth a budget of $1 mllon and a threshold of 50%, 85% of the parcels n the new threshold-constraned portfolo are also n the orgnal optmal portfolo derved wthout threshold constrants. When the threshold s ncreased to 90%, only 44% of the parcels n the optmal portfolo are also found n the orgnal portfolo. At a threshold of 50%, a larger budget of $2.5 mllon ncreases the overlap to 92%. There are, however, anomales, such as the greater overlap at a 90% threshold than at an 80% threshold under the nterval-scorng method and a $1 mllon budget. Such anomales can result because, as the threshold ncreases, the number of acqured parcels, n comparson to the orgnal, no-threshold portfolo, may ncrease or decrease non-monotoncally. To examne the effcency losses that arse when a conservaton agency gnores threshold constrants when acqurng contracts, we compare the scores of the portfolo that ncorporates threshold effects to the scores of the portfolo that gnores such effects. If the threshold constrant s not met n a catchment, contracts n that catchment yeld no water qualty benefts. The results are presented n Table 5. The effcency losses assocated wth gnorng thresholds are substantal, partcularly at low budget levels and hgh thresholds. For example, under a $1 mllon budget and an 80% threshold requrement, the portfolo derved wthout consderng the threshold constrants acheves zero benefts under three of the four scorng methods. A lower threshold at 50% mproves the portfolo s performance a lttle, but t stll acheves only 24% - 59% of what the portfolo derved under explct threshold constrants can acheve. Thus the data from Lake Skaneateles suggest that the falure to recognze nterdependent relatonshps among parcels that contrbute to achevng conservaton objectves can substantally 16
lower the effectveness of conservaton nvestments per dollar expended. Ignorng threshold effects, partcularly when avalable funds are few, can result n large foregone envronmental benefts. Interventons tend to be scattered over the landscape and fundng levels n any gven target area are nadequate to reach the threshold needed to mantan current water qualty levels or to acheve sgnfcant envronmental mprovements [Table 5 about here] The effcency losses are even more substantal when one compares the scores of the portfolo that recognzes threshold constrants and opportunty costs wth the scores of the portfolo that gnores threshold constrants and opportunty costs (.e., targetng on the bass of beneft scores alone). The results of ths comparson are presented n Table 6. Wth a budget of $1 mllon, the Cty of Syracuse would lkely generate no envronmental benefts f t were to acqure easements based on parcel scores alone. [Table 6 about here] Of course, the practtoner stll faces the problem of choosng among the dfferent optmal portfolos dentfed under each scorng rule. The practtoner could try the hgh-prorty approach of secton IV and focus on parcels that are found n the soluton of each scorng method, but the portfolos chosen through ths approach wll not necessarly acheve the thresholds n each catchment. In the Lake Skaneateles case, the hgh prorty portfolo of parcels selected from the optmal buffers when D = $2.5 mllon would come qute close to satsfyng the threshold requrements. In the 50% threshold scenaro, the hgh-prorty portfolo (cost = $1.52 mllon) spans 10 catchments, of whch 4 exceed the requred buffer-area threshold, 3 are less than 7% below the threshold, 2 are less than 19% below the threshold and 1 s less than 17
45% below the threshold. In the 80% threshold scenaro, the hgh-prorty portfolo (cost = $1.22 mllon) spans 5 catchments, of whch 2 exceed the threshold and 3 are less than 8% below the threshold. In the 90% threshold scenaro, the hgh-prorty portfolo (cost = $1.67 mllon) spans 4 catchments, of whch 2 exceed the threshold and 2 are less than 3% below the threshold. By ncreasng the budget or thresholds under whch the contract portfolos are chosen, a practtoner s more lkely to derve a hgh-prorty set of parcels that comes close to meetng the requred thresholds, although the degree to whch ths method s successful wll be case specfc. VII. If These Ideas Are So Great, Why Isn t Anyone Applyng Them? To many readers, the dea that ntegratng economc and bophyscal data can produce better results than smply usng bophyscal data alone seems straghtforward. If so, why s such an dea rarely appled n practce? As noted n Secton IV, there are condtons under whch acqurng and usng economc data wll not generate substantal mprovements n effcency. However, the emprcal studes cted n ths chapter mply that such condtons are not wdespread. Fve obstacles to ntegratng economc and bophyscal data are lkely to be much more mportant factors explanng why so few conservaton ntatves attempt to ncorporate bophyscal and economc data explctly n ther decson-makng. The frst and most obvous obstacle s the lack of awareness among conservaton practtoners about the basc concepts outlned above. Conservaton practtoners are rarely traned n economc theory and tend to read natural scence journals n whch few or no economsts publsh. In conservaton journals (e.g., Bologcal Conservaton), artcles on targetng conservaton nvestments contnue to be publshed wthout any reference to the opportunty costs of each nvestment. As noted n the ntroducton, academcs often do not take 18
nto account the objectves and approaches of practtoners and polcymakers. Furthermore, even wth an awareness of the basc concepts, practtoners may be unaware of the methods through whch bophyscal and economc data can be ntegrated. Although the smple methods presented n Secton III do not requre techncal tranng, more complcated analyses that ncorporate the nterdependent nature of landscape-level processes (e.g., bophyscal thresholds) requre more sophstcated methods of analyss n whch practtoners may lack tranng. Second, despte many publcatons on sophstcated bophyscal crtera-based conservaton targetng, practtoners have not adopted them. Prendergast et al. (1999) argue that practtoners often have a general antpathy toward what s seen as a prescrptve approach to conservaton.(p.484). Incorporatng economc data nto sophstcated targetng approaches only exacerbates the sense that the practtoner s flexblty to make decsons has been reduced. Based on conversatons wth practtoners, t seems as f practtoners often beleve they are mplctly ncorporatng economc costs nto the decson-makng and thus do not need to formally enter these costs nto the targetng algorthm. Thrd, obtanng relevant economc data can often be more dffcult than obtanng relevant bophyscal data because the latter s based on observable envronmental characterstcs whle the former s based on unobservable landowner characterstcs such as preferences. Practtoners often use two cost dscovery methods: (1) wat for a landowner to express nterest n a conservaton contract and then negotate over the contract prce (often the approach used by land trusts); or (2) estmate ex ante the lkely wllngness-to-accept of a small subset of landowners, typcally through real estate apprasal methods, and then negotate wth landowners sequentally by parcel rank. An ntegrated targetng approach requres that practtoners have reasonably accurate cost data for all parcels on whch one could potentally secure a contract. 19
Ths chapter used coarse apprasal data to proxy for the actual wllngness-to-accept of landowners, but t s uncertan how well such data represent the true underlyng dstrbuton of wllngness-to-accept. Practtoners already stretched to characterze bophyscal characterstcs on each parcel may fnd t too onerous to characterze the cost characterstcs ex ante. Fourth, conservaton practtoners who make the decsons on the ground are often judged not on the effcency wth whch they spend scarce conservaton dollars, but rather ther ablty to acheve gven objectves. For example, n the Lake Skaneateles ntatve and the New York Cty watershed ntatve, local agents are beng evaluated based largely on ther ablty to spend funds and meet acreage and farmer partcpaton levels. Platt et al. (2000) report that New York Cty s oblgated to commt $250 mllon durng the next 10 years to acqurng up to 335,000 acres. 7 The land conservaton portfolo that acheves these objectves may not be the same as the portfolo that meets an envronmental qualty objectve at least cost. The ffth, and most dffcult to tackle, obstacle les n the qualty of the data themselves. Many conservaton ntatves have budgets well below the budgets of large ctes lke Syracuse, New York Cty and Boston. In many of these ntatves, adequate bophyscal and economc data smply do not exst. When quanttatve data are poor, t s unclear whether any gans can be acheved through formalzaton of the decson-makng process. Indeed, such formalzaton mght lead to egregous errors. The use of subjectve expert opnon to gude decson-makng mght be more approprate n such cases. VII. Concluson 7 Under the agreement t has wth the EPA, New York Cty efforts must also entce at least 85 percent of the farmers n the watershed to jon the polluton reducton program. 20
Polcymakers and conservaton practtoners throughout the world seek flexble tools that permt the ntegraton of bophyscal and economc data nto cost-effectve conservaton plans. Ths chapter demonstrates a way n whch conservaton agences can ntegrate spatally varable bophyscal and economc data n the absence of sophstcated bophyscal modelng. Usng common bophyscal scorng methods, n combnaton wth economc data and smple optmzaton methods, a set of prorty land parcels can be dentfed for contractng. In an emprcal applcaton, data from a Geographc Informaton System (GIS) are used to dentfy a set of prorty land parcels for a rparan buffer contractng ntatve n upstate New York. To ensure that the results from ths applcaton were useful, the data selected for ths applcaton came from the decson-makers themselves. Furthermore, the analyss explctly consders the methods beng used by decson-makers and approaches the problem at the geographc scale at whch decsons are beng made. Ths chapter also demonstrates a way n whch conservaton agences can ncorporate concerns about bophyscal thresholds n ther decsonmakng. The results corroborate prevous emprcal work suggestng that the falure to consder economc data n envronmental nvestment decsons can lead to large losses n effcency. Moreover, fndngs reveal that the potental effcency losses assocated wth gnorng bologcal thresholds are also large. In Secton VI, reasons why practtoners may not desre or be able to ntegrate bophyscal and economc data n ther decson-makng are explored. The actual decson process s emphaszed n ths chapter rather than the bophyscal modelng, but clearly the results are only as good as the bophyscal and economc nformaton on whch the analyss s based. We take as gven the data avalable to the Cty of Syracuse and the way n whch the Cty s practtoners express ther preferences and objectves. However, f the relablty of the parcel-scorng functons or the threshold estmates s poor, there s no 21
guarantee the tools developed n ths chapter mprove upon current practtoner methods. 8 The same caveat holds for the estmates of contractng costs. The use of hgh-prorty portfolos, lke those dentfed here, may mtgate errors n beneft and cost estmaton, but scholars and practtoners need to ensure that they have relable nformaton to feed nto the decson analyss. Integratng relable bophyscal and economc nformaton s partcularly mportant n the context of watershed conservaton for three reasons: (1) the level of envronmental amentes and the costs of obtanng the amentes are lkely to be postvely correlated (e.g., conservaton on large parcels wth extensve waterfront and located near nfrastructure are lkely to be mportant for water qualty objectves, but are also lkely to be expensve), (2) n rapdly developng watersheds, the relatve spatal varablty of conservaton contract costs s lkely to be greater than the relatve spatal varablty of conservaton benefts, and (3) uncoordnated efforts to establsh rparan buffers across the watershed are lkely to lead to lttle or no water qualty benefts. Collectvely, these factors suggest that f practtoners fal to ntegrate the avalable bophyscal and economc data, currently popular conservaton contractng approaches for watershed protecton may acheve far fewer envronmental benefts than expected. References Allen, A.W. 1983. Habtat Sutablty Index Models: Beaver. US Fsh and Wldlfe Servce publcaton number FWS/OBS-82/10.30 Revsed. Fsh and Wldlfe Servce, Washngton, DC. Allen, A. W., and R. D. Hoffman. 1984. Habtat sutablty ndex models: Muskrat. U. S. Fsh and Wldlfe Servce. FWS/OBS-82/10.46. Fsh and Wldlfe Servce, Washngton, DC. Ando, A, J. Camm, S. Polasky, and A. Solow. 1998. Speces dstrbutons, land values, and effcent conservaton. Scence 279: 2126-2128. 8 Although the use of scorng functons lke those used n ths paper s wdespread, there s evdence that lnear preference functons may be a poor proxy for decson-maker preferences (Keeney and Raffa, 1976) and that the dentfcaton of crtera weghts s complcated even for experts (Borcherdng et al., 1993). See Ferraro (2004) for an alternatve. 22
Azzano, Z., J.M. Conrad and P.J. Ferraro. 2002. Optmzng the Rparan Buffer: Harold Brook n the Skaneateles Lake Watershed, New York. Land Economcs 78(4): 501-514. Babcock, B.A., P.G. Lakshmnarayan, J. Wu, and D. Zlberman. 1997. Targetng Tools for the Purchase of Envronmental Amentes. Land Economcs 73(3): 325-39. Babcock, B.A., P.G. Lakshmnarayan, J. Wu, and D. Zlberman. 1996. The Economcs of a Publc Fund for Envronmental Amentes: a study of CRP contracts. Amercan Journal of Agrcultural Economcs 78: 961-971. Balmford, A., K.J. Gaston, and A.S.L. Rodrgues. 2000. Integratng costs of conservaton nto nternatonal prorty settng. Conservaton Bology 14(3): 597 604. Balmford, A., K.J. Gaston, S. Blyth, A. James, and V. Kapos. 2003. Global varaton n conservaton costs, conservaton benefts, and unmet conservaton needs. Proceedngs of the Natonal Academy of Scences 100: 1046-1050. Borcherdng, K., Schmeer, S., and Weber, M. 1993. Bases n Multattrbute Weght Elctaton, n J.P. Cavern, M. Bar-Hllel, F.N. Barron and H. Jungerman, eds. Contrbutons to Decson Research. North Holland, Amsterdam, The Netherlands, pp. 3-28. Bulte, E. H. and G.C. van Kooten. 2001. Harvestng and Conservng a Speces when Numbers Are Low: Populaton Vablty and Gambler's Run n Boeconomc Models. Ecologcal Economcs 37(1): 87-10. Carter, M.F., W.C. Hunter, D.N. Pashley, and K.V. Rosenberg. 1999. Settng Prortes Prortes for Landbrds n the Unted States: the Partners n Flght Approach. Ithaca, NY, Cornell Laboratory of Ornthology. Cummngs, R.G., S.K. Laury and C. Holt. 2003. The Georga Irrgaton Reducton Aucton: Experments and Implementaton. Journal of Polcy Analyss and Management 23(2): 341-363. DeZyane, J. 1990. Handbook of Drnkng Water Qualty. John Wley & Sons, New York, NY. Dobson, A.P., J.P. Rodrguez, W.M. Roberts, D.S. Wlcove. 1997. Geographc dstrbuton of endangered speces n the Unted States. Scence 275(January 24): 550 555. Envronmental Protecton Agency (EPA), Offce of Enforcement and Complance Assurance (OECA), Offce of Federal Actvtes. 1999. Consderaton of Cumulatve Impacts n EPA Revew of NEPA Documents. May. Washngton, D.C. http://es.epa.gov/oeca/ofa/cumula.html Farzn, Y. H. 1996. Optmal Prcng of Envronmental and Natural Resource Use wth Stock Externaltes. Journal of Publc Economcs 62(1-2): 31-57. 23
Ferraro, P.J. 2004. Targetng Conservaton Investments n Heterogeneous Landscapes : A dstance functon approach and applcaton to watershed management. Amercan Journal of Agrcultural Economcs 86(4): 905-918. Ferraro, P.J. 2003. Assgnng Prorty to Envronmental Polcy Interventons n a Heterogeneous World. Journal of Polcy Analyss and Management 22(1): 27-43. Ferraro, P.J. 2002. Conservaton Contractng n Heterogeneous Landscapes: an applcaton to watershed protecton wth threshold constrants. Georga State Unversty Water Polcy Workng Paper Seres 2002-010. Atlanta, GA. Avalable at http://epp.gsu.edu/pferraro/research/workngpaper/workngpapers.htm. Ferraro, P.J., and A. Kss. 2002. Drect Payments for Bodversty Conservaton. Scence 298 (29 November): 1718-1719. Ferraro, P.J. 2001. Water Qualty Protecton and the Cost-effectve Targetng of Rparan Buffers n Georga. GSU Water Polcy Workng Paper #2001-004. Dept. of Economcs, Andrew Young School of Polcy Studes. Atlanta, GA. Feather, P., D. Hellersten, and L. Hansen. 1998. Explorng Methods of Selectng Cropland for Conservaton. Agrcultural Outlook (September): 21-24. Florda Department of Envronmental Protecton (FDEP), Surface Water Montorng Secton. 2000. Current Bologcal Health and Water Qualty of the Econlockhatchee Rver and Selected Trbutares. January and July 1999. FDEP, Orlando, Florda. Gardner, K.V. 2000. Skaneateles Lake Watershed Land Protecton Program: methodology and market analyss for conservaton easement valuaton. North East Apprasals and Management Co., Inc, Ithaca, NY. Hermans, L. M. 1999. Desgn of Phosphorus Management Strateges for the Cannonsvlle Basn. New York State Water Resources Insttute, Cornell Unversty, Ithaca, New York. Hruby, T, S. Stanley, T. Granger, T. Duebendorfer, R. Fresz, B. Lang, B. Leonard, K. March, and A. Wald. 2000. Methods for Assessng Wetland Functons Volume II: Depressonal Wetlands n the Columba Basn of Eastern Washngton. Washngton State Department Ecology Publcaton #00-06-47. Pullman, WA. Johnson, N., C. Revenga and J. Echeverra. 2001. Managng Water for People and Nature. Scence 292: 1071-1072. Keeney, R.L., and Raffa, H. 1976. Decsons wth Multple Objectves: preferences and value tradeoffs. John Wley, New York, NY Lande, R. 1987. Extncton thresholds n demographc models of terrtoral populatons. The Amercan Naturalst (13): 624-635. 24
Lemunyon, J.L. and R.G. Glbert. 1993. The Concept and Need for a Phosphorus Assessment Tool. Journal of Producton Agrculture 6: 483-486. Master, L. 1991. Assessng Threats and Settng Prortes for Conservaton. Conservaton Bology 5: 559-563. McMahon, T.E. 1983. Habtat Sutablty Index Models: Coho Salmon. FWS/OBS-82/10.49. USDI Fsh and Wldlfe Servce, Habtat Evaluaton Procedure Group, Ft. Collns, CO. Metropoltan Dstrct Commsson (MDC), Dvson of Watershed Management. 1999. Land Acquston Fact Sheet, FS98.01, July. Boston, MA. http://www.state.ma.us/mdc/landacq.pdf Mttermer, R.A., N. Myers, J.B. Thompsen, G.A.B. Fonseca, and S. Olver. 1998. Global bodversty hotpsots and major tropcal wlderness areas. Conservaton Bology 12: 516-520. Myers, S., L. Macbeth, and R. Nemecek. 1998. Skaneateles Lake Watershed Management Plan. Prepared for the U.S. Envronmental Protecton Agency and The New York State Department of Envronmental Conservaton. Cty of Syracuse, New York. Natonal Research Councl (NRC), Commttee to Revew the New York Cty Watershed Management Strategy. 2000. Watershed Management for Potable Water Supply: assessng New York Cty s Approach. Natonal Academy Press, Washngton, D.C. New York State Department of Publc Health. 1999. New York State Source Water Assessment Program Plan. New York State Department of Publc Health, Bureau of Publc Water Supply Protecton, 547 Rver Street, Flangan Square, Room 400, Troy, New York. Olson, D. M., E. Dnersten, R. Abell, T. Allnutt, C. Carpenter, L. McClenachan, J. D Amco, P. Hurley, K. Kassem, H. Strand, M. Taye, and M. Theme. 2000. The Global 200: a representaton approach to conservng the Earth s dstnctve ecoregons. Conservaton Scence Program, World Wldlfe Fund, Washngton, D.C. Platt, R.H., P.K. Barter and M.J. Pfeffer. 2000. A Full, Clean Glass? Managng New York Cty s Watersheds. Envronment 42(5): 8-20. Polasky, S., J.D. Camm, and B. Garber-Yonts. 2001. Selectng Bologcal Reserves Costeffectvely: an applcaton to terrestral vertebrate conservaton n Oregon. Land Economcs 77(1): 68-78. Prendergast, J.R., R.M Qunn, and J.H. Lawton. 1999. The Gaps between Theory and Practce n Selectng Nature Reserves. Conservaton Bology 13(3): 484-492. Rowles, J.L., and D. L. Stlnger. 1999. Assessment of Interstate Streams n the Susquehana Rver Basn. Montorng Report No. 12. July 1, 1997, to June 30, 1998. Publcaton no. 205. Water Qualty and Montorng Program Susquehanna Rver Basn Commsson. Harrsburg, PA. 25
Schueler, Thomas. 1994. The Importance of Impervousness. Watershed Protecton Technques 1 (Fall): 100. Schueler, Thomas. 1995. The Peculartes of Pervousness. Watershed Protecton Technques 2 (Fall): 233. Shaffer, M. L. 1981. Mnmum populaton szes for speces conservaton. BoScence 31: 131-134. Smth, R.D., A. Ammann, C. Bartoldus, and M.M. Brnson. 1995. An Approach for Assessng Wetland Functons Usng Hydrogeomorphc Classfcaton, Reference Wetlands, and Functonal Indces. US Army Engneers Waterways Experment Staton, Techncal Report WRP-DE-10, and Operatonal Draft, Vcksburg, MS. Terrell, J.W., T.E. McMahon, P.D. Inskp, R.F. Ralegh, and K.L. Wllamson. 1982. Habtat sutablty ndex models: Appendx A. Gudelnes for rverne and lacustrne applcatons of fsh HSI models wth the Habtat Evaluaton Procedures. U.S. Department of the Interor, Fsh and Wldlfe Servce FWS/OBS-82/10.A. Tlman, D., J. Fargone, B. Wolff, C. D Antono, A. Dobson, W. Howarth, D. Schndler, W.H. Schlesnger, D. Smberloff, and S. Swackhamer. 2001. Forecastng Agrculturally Drve Global Envronmental Change. Scence 292(13 Aprl): 281-284. Unted States Fsh and Wldlfe Servce (USFWS). 1981. Standards for the development of habtat sutablty ndex models. 103 ESM. US Fsh and Wldlfe Servce, Washngton, DC. Unted States Department of Agrculture (USDA), Farm Servce Agency. 1999. Conservaton Reserve Program Sgn-up 20: Envronmental Benefts Index. Fact Sheet. Farm Servce Agency, Washngton, DC. Voogd, H. 1983. Multcrtera Evaluaton for Urban and Regonal Plannng. Pon Lmted, London, UK. Wang, L., J. Lyons, P. Kanehl, R. Banneman, and E. Emmons. 2000. Watershed Urbanzaton and Changes n Fsh Communtes n Southeastern Wsconsn Streams. Journal of the Amercan Water Resources Assocaton. 36(5): 1173-1190. Wang, L., J. Lyons, P. Kanehl, and R. Gatt. 1997. Influences of Watershed Land Use on Habtat Qualty and Botc Integrty n Wsconsn Streams. Fsheres 22(6): 6-12. Wu, J., R.M. Adams, W.G. Boggess. 2000. Cumulatve Effects and Optmal Targetng of Conservaton Efforts: Steelhead Trout Habtat Enhancement n Oregon. Amercan Journal of Agrcultural Economcs 82(2): 400-413. Zoner, D., and N.M. Lmtz. 1994. Should Numercal Impervousness Be Used to Zone Watersheds? Center for Watershed Protecton. Watershed Protecton Technques (Summer)1: 2. 26