Effcet Compesato for Regulatory Takgs ad Orego s Measure 37 Jack Scheffer Ph.D. Studet Dept. of Agrcultural, Evrometal ad Developmet Ecoomcs The Oho State Uversty 2120 Fyffe Road Columbus, OH 43210-1067 scheffer.1@osu.edu Tel. (614) 292-9403 Fax (614) 292-7710 Selected Paper prepared for presetato at the Amerca Agrcultural Ecoomcs Assocato Aual Meetg, Portlad, OR, July 29-August 1, 2007 Ths verso: July 27, 2007 Copyrght 2007 by Jack Scheffer. All rghts reserved. Readers may make verbatm copes of ths documet for o-commercal purposes by ay meas, provded that ths copyrght otce appears o all such copes.
I regulatg lad-use, govermets must balace socetal terests evrometal protecto, cotrol of urba sprawl, ad smlar ssues wth the terests of prvate owers developg ad usg lad. Sce restrctg lad use order to meet publc goals ofte etals prvate losses decreased market value of lad or foregoe profts the questo arses of whether property owers should be compesated for these losses, ofte called regulatory takgs. Compesato requremets, such as the recetly eacted Measure 37 Orego, may be advocated for reasos of both faress ad ecoomc effcecy. Ths artcle exames the effcecy aspect of full-compesato rules, whch requre regulators to pay property owers for losses due to regulato. Prevous aalyses have geerally foud that full-compesato rules lead to effcet outcomes the form of overvestmet by property owers. However, gve the cotued poltcal support for such rules, oe aturally woders whether they mght be justfed some settgs. I ths artcle, I frst develop a model showg oe set of codtos uder whch the overvestmet problem s elmated ad a fullcompesato rule ca therefore be justfed o effcecy grouds. Ths result provdes a possble postve explaato for why such rules are advocated. It also provdes a ormatve prescrpto for mprovg the effcecy of such rules whe they are desrable for other reasos (e.g. smplcty of mplemetato, faress). Fally, ths result provdes a bechmark for cosderg the effcecy of full compesato whe I exted the model to clude budget costrats ad taxato, two mportat real-world cosderatos largely abset from prevous aalyses. Usg ths exteded model, I show that, whe 1
govermets face budget costrats ad lmted powers of taxato, a full-compesato rule leads to effcecy the form of uder-regulato. Lterature Revew Ies, Polasky, ad Tschrhart (1998) survey the ecoomcs lterature o regulatory takgs ad the effcecy effects of compesato schemes. Much of ths lterature aalyzes whether cetve schemes wll lead to effcet choces by two agets: prvate property owers ad govermet regulators. I partcular, t studes whether rules requrg moetary compesato promote optmal decsos at two juctures. Frst, the rule should ecourage the property ower to make the socally optmal choce for the use of the lad, cludg ay vestmet t. For example, the ower may choose amog growg tmber for harvest, clearg the lad for crops, ad subdvdg the property for resdetal developmet. These alteratve uses may have dfferet tal vestmet costs for the ower, whch may be urecoverable f he s later forced to swtch to a dfferet use. Also, these alteratves wll lkely offer dfferet potetal returs to the ower ad cur dfferet exteral costs (or beefts) for socety. Ecoomc models geerally assume that the property ower wll cosder oly hs prvate costs ad beefts whe makg the lad use decso. If there are sgfcat exteraltes assocated wth some of the uses, hs proft-maxmzg decso may ot lead to the optmal choce from a socal perspectve. Thus, maxmzg socal welfare may ecesstate regulato or prohbto of certa lad uses. A effcet compesato rule wll combe wth regulato to esure socally optmal decsos by the lad ower. 2
Secod, the rule should ecourage effcet regulatory decsos by the govermet, such as determg whether to allow a partcular lad use for ay gve parcel. I geeral, a govermet s ofte assumed to employ the Kaldor-Hcks (.e. potetal Pareto mprovemet) crtero ts regulato decsos, leadg to effcet regulato. I the cotext of takgs ad regulatory takgs, however, observers have log feared that the govermet may ot always act socety s best terest wthout proper cetves. Such cocers cotrbute to support for compesato requremets, such as the Ffth Amedmet takgs clause: or shall prvate property be take for publc use, wthout just compesato. Ths ratoale for the takgs clause s a moral hazard argumet: the govermet wll ted to exercse emet doma too ofte f takgs are costless to t. Extedg ths argumet to regulatory takgs, some fear that the govermet wll gore the costs of regulato to property owers, leadg to overregulato. Ths fscal lluso, as t s called the lterature, would requre some costrat o regulatory behavor, such as a compesato requremet, to brg the regulator s decso to algmet wth maxmzg aggregate socal welfare. Ecoomsts have vestgated varous compesato rules that attempt to balace these two dstortos lad owers dsregard for exteraltes ad govermet s fscal lluso ad lead to socally optmal decsos by all partes. I a semal artcle, Blume, Rubste, ad Shapro (1984) foud that f the ower s compesated for lost market value, he wll have a cetve to over-vest (e.g. choose lad uses that requre upfrot vestmets that are wasted whe regulato s eacted). A full-compesato rule essetally sures the ower agast the rsk of regulato. Ths model suggests that a o- 3
compesato rule 1 would lead the ower to make the socally optmal choce. Ths result assumes that govermet takgs are exogeous to the level of vestmet ad are presumed effcet. (I.e. Fscal lluso s ot preset.) Mcel ad Segerso (1994, 1996) develop a model to provde a postve explaato for the commo law approach to regulatory takgs. I ther model, the takgs decso s edogeous ad subject to fscal lluso. As dd Blume et al, they fd that full compesato leads to effcecy, as the ower wll ted to over-vest. They fd two effcet rules that feature compesato codtoed o effcecy, echog Supreme Court decsos that allow compesato oly whe a regulato goes too far. Ther ex ate rule provdes full compesato to the lad ower f ad oly f hs vestmet decso was ex ate effcet (.e. wth regard to expectatos about exteraltes). The effcecy of the govermet s decso s rrelevat for determg compesato. Uder ther ex post rule, alteratvely, full compesato s pad f ad oly f the govermet s regulatory decso was ex post effcet (.e. gve the ower s vestmet choce ad the realzed exteralty). Uder ether rule, both the property ower ad the govermet should make effcet choces. Hermal (1995) vestgates effcet compesato for edogeous takgs uder several formato structures. He fds that full compesato (.e. lost market value) s geerally effcet, leadg to overvestmet due to the ower s moral hazard. He also fds that a zero-compesato rule s effcet. Uder such a rule, the ower has a cetve to over-vest order to dscourage govermet takgs. The 1 The Blume, Rubfeld, ad Shapro result requres lump-sum compesato, whch does ot vary wth the vestmet level. Zero compesato, the best-kow verso of the rule, s a specal case of ths result. 4
dfferece betwee ths result ad the Blume et al result stems from the edogeety of the takgs decso Hermal s model, whereas Blume et al assume exogeous takgs. 2 Hermal shows that whe fscal lluso does ot exst, ether because the govermet s beevolet or because the amout of the exteralty s commo kowledge, the effcet compesato rule pays the prvate ower the socal beeft the evet of a takg. As Ies et al (1998) pot out, ths rule s essetally Pgouva compesato. Whe the state s ot beevolet ad the exteralty s the state s prvate formato, Hermal s effcet rule volves two-part compesato. Frst, the state aouces ts terest a takg ad asks the ower to state a prce reflectg hs prvate value. Ths aoucemet trggers a trasfer equal to the state s expected surplus f the takg proceeds at the ower s stated prce. Ths mechasm duces the ower to truthfully reveal hs prvate value. The state the has the opto to cotue wth the takg, whch requres a further paymet at the ower s stated prce. Ies (1997) troduces multple parcels ad a tme dmeso, to vestgate the effects of the relatve treatmet of developed ad udeveloped parcels. He fds that the compesato rule has a sgfcat effect o the tmg of developmet. I partcular, a full-compesato rule duces owers to develop too early, order to ga the beefcal treatmet that developed propertes receve a takg uder such a rule. He further fds that over-compesatg owers of udeveloped lad would restore the 2 The exogeous takg assumpto s made the basc model (Secto II) Blume et al (1984), resultg the zero-compesato rule. I ther Secto III, the authors cosder edogeously determed takgs, stll wthout fscal lluso, ad fd that both zero- ad full-compesatos rules duce effcecy the form of overvestmet, parallelg Hermal s (1995) later result. 5
equalty of cetves ecessary for effcet tmg of developmet. He the advocates the use of tradable developmet rghts as a alteratve to moetary compesato. Fscal Illuso The assumpto of fscal lluso plays a key role most argumets for a compesato rule of some form. The usual arratve s that the govermet s more resposve to oe segmet of socety (e.g. coservatosts) tha to others (e.g. lad owers). Thus, t bases ts regulatory decsos dsproportoately o the frst group s costs ad beefts, gorg or dscoutg the other groups terests. Buchaa (1962) argues that t s cosstet for ecoomc models to assume that dvduals are motvated by self-terest ther persoal affars whle also assumg that govermet decso-makers pursue oly maxmum socal welfare performg ther jobs. The theores of the publc choce feld stead use models of self-terested govermet actors, motvated by re-electo, bureaucratc empre-buldg, ad smlar objectves. I such a settg, a regulator s weghtg of oe group s terests over aother s seems qute reasoable. Oe group may possess more poltcal clout tha aother. A regulatory agecy may push for more regulato, to justfy ts exstece ad exteds ts sphere of fluece. Thus, govermet agets, ratoally pursug ther ow self-terests, lead govermet to devate from the goal of maxmzg socetal welfare. However, Ies (1997) argues that fscal lluso may be a urealstc assumpto. He develops a model of the poltcal behavor of a ufettered govermet oe free to set both taxato ad regulato that respods more to oe terest group tha to aother. Ies fds that the govermet s best served by usg effcet 6
regulato, thus maxmzg the surplus that t has avalable to tax, eve whe t favors oe group over aother. However, f a govermet taxato s ad redstrbuto powers are ot ufettered, fscal lluso may stll exst. Suppose the govermet s uable to tax the full surplus of the dsfavored group. Usg regulato to trasfer surplus from the dsfavored group to the favored group may yeld a greater beeft to the latter tha would taxato, eve f some loss occurs from the effcecy of regulato. I the exteded model, I corporate both a lmted power to tax ad a budget costrat that lmts the govermet s ablty to pay compesato. Uder these sttutoal costrats, fscal lluso ca stll exst. Basc Model I ths secto, I develop a model wth whch I wll compare compesato rules uder two possble sequeces of evets. Scearo 1 follows the sequece used much of the prevous lterature, ad the model replcates the famlar result that a full-compesato rule s effcet. Scearo 2 presets a alteratve sttutoal framework, uder whch a full-compesato rule performs more favorably. A parcel of lad has two potetal uses: developed ad udeveloped. Developmet requres a upfrot vestmet, I > 0, ad curs a potetal exteral cost to socety the form of polluto, lost evrometal servces, reducto aesthetc value, or smlar harm. The amout of the exteral cost s a radom varable, s. Udeveloped property requres o vestmet ad curs o exteral cost to socety. The prvate value of udeveloped property to the ower s ormalzed to 0. The prvate value of developed property s a radom varable, p, wth a expected value of p. 7
The govermet may choose to regulate lad use o the parcel, prevetg developmet. If regulato s mposed, ay upfrot vestmet I s lost, the ower receves a prvate beeft of 0, ad socety avods exteral harm of s. The govermet may also have to pay compesato, C, to the property ower, depedg upo the compesato rule place. Varables p ad s have desty fuctos f(p) ad g(s) wth cdfs F(p) ad G(s), respectvely, ad p ad s are depedetly dstrbuted. I order to smplfy the problem ad restrct atteto to terestg cases, assume that F(I) = 0, so that developmet s always prvately optmal. Also assume G(0) = 0, so that the realzed socal cost of developmet wll always be postve. I assume that both the property ower ad the regulator are rsk-eutral. The sequece of evets wll follow oe of two scearos. Scearo 1 follows the patter establshed the prevous lterature, ad the model replcates the famlar results cocerg compesato rules. Scearo 2 presets a alteratve framework, whch dfferet results follow. I partcular, a full-compesato rule performs favorably. A dscusso of whch scearo best reflects the ssues facg polcy-makers follows. Scearo 1 Uder scearo 1, the sequece of evets s as follows. At tme 1, the ower chooses a lad use, payg I f developmet s chose. At tme 2, the radom varables p ad s are realzed, ad the govermet chooses whether to regulate lad use o the parcel. At tme 3, ay requred compesato s pad. 8
Frst-Best Soluto A hypothetcal socal plaer seeks to maxmze expected aggregate welfare: prvate beeft less vestmet ad exteral socal cost. Ths plaer makes decsos place of the prvate ower at tme 1 ad place of the regulator at tme 2. PROPOSITION 1: Uder scearo 1, the socal plaer would vest for developmet at tme 1 f ad oly f gs f p p sdpds I s ad would allow developmet to proceed at tme 2 f ad oly f p s > 0. (Proofs for all propostos are preseted the appedx.) The hypothetcal socal plaer s decso rules at tmes 1 ad 2 provde a bechmark for aalyzg the effcecy of the decsos the prvate ower ad the govermet regulator would make uder varous compesato rules. Regulator s Problem At tme 2, the decso whether to proceed wth developmet (assumg the vestmet was made) falls to the govermet regulator. The regulator s assumed to operate uder fscal lluso, essetally gorg the prvate ower s terests. 3 The regulator stead acts to maxmze the welfare of the rest of socety, subject to ay requred compesato. Whe the prvate ower has vested for developmet, the regulator s decso problem ca be expressed as Max 1 rs rc r, where r = 1 whe the govermet chooses to regulate, dsallowg developmet, ad r = 0 otherwse. The form ad value of C wll deped o the compesato rule place. Uder a full-compesato rule, C = p, 3 A less restrctve verso of fscal lluso would allow dfferet, but o-zero, weghts for the terests of the ower ad the rest of socety. Blume et al (1984) preset such a model ther Secto V. I order to reta smplcty, I use a mplct weght of zero for the ower s welfare. 9
because the govermet must pay the prvate ower the lost value of developmet that s precluded by the regulato. Uder a o-compesato rule, C = 0, because the govermet pays o compesato regardless of whether t regulates. PROPOSITION 2: Uder a o-compesato rule scearo 1, the regulator wll prevet developmet f ad oly f s > 0. Uder a fullcompesato rule scearo 1, the regulator wll allow developmet f ad oly f p s > 0. Uder the o-compesato rule, the govermet always regulates, sce s > 0 s true by assumpto. Ths amouts to over-regulato wheever p > s s realzed, because the socally optmal outcome s o regulato that case. Uder ths rule, the regulator ever uder-regulates (.e. chooses r = 0 whe s > p). The full-compesato rule, cotrast, teralzes the prvate ower s value ad duces a decso rule equvalet to the socal plaer s rule. Thus, the regulator s objectve fucto uder full compesato s alged wth maxmzg aggregate welfare. Ower s Problem At tme 1, the prvate ower s faced wth the choce of whether to make the upfrot vestmet that wll allow developmet. Sce the prvate beeft ad socal cost of developmet are ot yet realzed, he makes hs vestmet choce based o expectatos of hs prvate beefts ad the socal cost, the regulator s decso rule, the compesato rule, ad hs vestmet cost. Hs problem ca be expressed as de1 r Max p rc I, where d = 1 f he makes the vestmet for developmet ad d = 0 otherwse. Aga, the form of C depeds o the compesato rule effect. d 10
PROPOSITION 3: Uder scearo 1, a ower facg a o-compesato rule ever vests for developmet. A ower facg a full-compesato rule always vests for developmet. Thus, a o-compesato rule leads the ower to uder-vest respose to the expectato that the regulator wll dsallow developmet, eve whe t would be effcet. The full-compesato rule duces the ower to over-vest, choosg to vest I all cases, eve though effcecy requres that he should forego vestmet some cases, gve the possblty that regulato wll be chose. Summary I scearo 1, a full-compesato rule leads to effcet regulato but effcet overvestmet, a result cosstet wth the prevous lterature. A o-compesato rule leads to overregulato ad correspodg udervestmet as a respose. More complcated compesato rules, such as those of Mcel ad Segerso (1994, 1996) ad Hermal (1995), would be ecessary to acheve effcecy uder these codtos. The key elemet of the scearo s that the property ower chooses vestmet before the govermet has the opportuty to assess socal cost ad regulate lad use. Scearo 2 Suppose stead that evets occur the followg sequece. At tme 1, the govermet mplemets a permt requremet, requrg the ower to apply for a permt before developg hs property. At tme 2, radom varable p s realzed ad the ower submts a permt applcato, specfyg p ad I. At tme 3, radom varable s s realzed, ad the govermet chooses whether to allow developmet. At tme 4, ay requred 11
compesato s pad or vestmet ad developmet proceed f permtted. All other elemets of the model rema uchaged. Ths sequece correspods to a evromet whch a system of regulatos s already place. Whe a ew developmet opportuty arses, the lad ower s requred to apply for approval from the approprate regulatory agecy before vestg for developmet. If the agecy choose to eforce the regulato, the compesato may be requred. If the regulato s waved for that property, developmet proceeds ad o compesato s due. Frst-Best Soluto The socal plaer maxmzg expected aggregate welfare has oly oe choce to make, at tme 3, whether to allow vestmet ad developmet. PROPOSITION 4: Uder scearo 2, the socal plaer wll proceed wth developmet at tme 3 f ad oly f p I > s. Ths decso rule maxmzes aggregate welfare, because t allows developmet exactly whe the aggregate beefts outwegh the aggregate costs, cludg upfrot vestmet. Regulator s Problem As scearo 1, the regulator determes whether to allow developmet to proceed, at tme 3 uder ths scearo. The assumpto of fscal lluso s mataed, so the regulator effectvely gores the prvate ower s terest uless the compesato rule makes t relevat. PROPOSITION 5: Uder scearo 2, a govermet facg a fullcompesato rule wll allow developmet at tme 3 f ad oly f p I > s. 12
Thus, the govermet s regulato decso s effcet, replcatg that of the socal plaer. Ower s Problem I scearo 2, the ower s problem s trval. He wll always develop at tme 4 f permtted, but developmet wll be permtted oly whe t s effcet. Thus, the ower s decso at ths stage also coforms to the socal plaer s rule ad s effcet. At tme 2, the ower submts a applcato dcatg hs desre to develop lad, as well as formato o p ad I. For the sake of smplcty, ths aalyss assumes truthful revelato by the ower at ths stage. However, f p ad I are prvate formato, the potetal for dshoest applcatos exsts. The ower may overstate the proft potetal from developmet, p I, partcularly a o-compesato settg. 4 Eve uder a full-compesato settg, the ower may overstate p I order to crease hs potetal compesato, although dog so decreases the probablty that regulato wll fact be eforced. Of course, the state may requre evdece to substatate the ower s clam, order to mtgate the potetal for dshoesty. A full aalyss of the formato ad verfablty aspects of ths process s beyod the scope of ths artcle. Summary I scearo 2, a full-compesato rule leads to effcet decsos. Because the ower s preveted from vestg utl the regulato choce s made, vestmet wll be effcet. The full-compesato rule teralzes the ower s prvate costs (lost beeft) the regulator s objectve fucto, leadg to effcet regulatory decsos as well. 4 Uder a o-compesato rule, revelato of p I s relevat oly f the regulator s assumed to place some weght o the prvate beefts of the lad ower (.e. partal fscal lluso). 13
Whch Scearo Is More Realstc? The U.S. commo law has geerally draw a dstcto betwee physcal takgs (e.g. usg emet doma to seze lad for buldg a hghway) ad regulatory takgs (e.g. a prohbto agast developmet whle the owers reta ttle to the lad). For a physcal takg, the govermet has bee oblgated uder the Ffth Amedmet to pay just compesato to the ower, usually terpreted as far market value for the property. For regulatory takgs, the Supreme Court has artculated several crtera (see Mcel ad Segerso 1994, 1996 for a descrpto) for determg uder what crcumstaces the govermet must pay compesato. However, the stadard s far from clear ad, practce, compesato s ot foud to be requred the vast majorty of court cases. I cotrast to the commo law, the ecoomcs lterature has geerally treated physcal ad regulatory takgs as a sgle ssue, varyg the proporto of property rghts take but smlar eough coceptually for the same theoretcal models to apply. The early lterature focused o physcal takgs ad developed models reflectg the sequece scearo 1. Ths sequece seems very reasoable the case of physcal takgs, sce ay relevat actos by the ower must occur before the takg; he wll have lttle opportuty for vestmet oce he loses ttle. Subsequet lterature bega to address regulatory takgs, although ofte usg models derved from aalyses of physcal takgs. Thus, the sequece from scearo 1 forms the bass of these models. However, t may be that scearo 2 represets a better approxmato of realty, at least some regulatory stuatos. Orego s Measure 37, for example, follows the sequece of scearo 2. Uder that law, log-stadg regulatos of 14
varous sorts preclude developmet uless owers apply for ad receve a waver of the regulato. From a prescrptve perspectve, scearo 2 offers gudeles for mplemetg a full-compesato rule a more effcet way f such a rule s desrable for reasos beyod the effcecy crtero. Such a rule mght be appealg because of ts relatve smplcty or because t s cosstet wth otos of procedural faress (e.g. spreadg the costs of socetal beefts wdely amog taxpayers). I such a case, desgg mechasms to prevet vestmet lad uses before a fal regulatory decso s made may avod some effcecy lad owers choces, whle retag the effcet cetves a full-compesato rule provdes for regulators. For purposes of ths artcle, scearo 2 provdes a desrable bechmark. Uder scearo 2, a full-compesato rule acheves the effcet outcome the basc model, whch omts budget ad taxato ssues. By examg scearo 2 uder a exteded model corporatg those elemets, I ca solate the effcecy mpacts of those elemets wth regard to a full-compesato rule. Budget Costrats ad Taxato A frequet crtcsm of Measure 37 s that the law s a de facto repeal of the state s comprehesve lad use plas. These crtcs clam that, sce the state ageces have lttle moey for payg compesato clams, they wll be forced to wave the regulatory restrctos may cases. There seems to be some emprcal support for ths crtcsm. As of Jauary 2006, o resoluto of ay clam had compesato bee pad; that s, wavers had bee chose for all legtmate clams (Mart ad Shrver, 2006). 15
Budget ad taxato ssues are realstc cocers for a aalyss of govermet decso-makg. For example, the budget of the govermet as a whole s costraed by ts taxato power; t caot sped more moey tha t ca rase va taxes. 5 Idvdual ageces may ot possess the power to tax, ad therefore rely o appropratos to fud ther spedg. These factors support the relevace of a budget costrat. Govermet may be preveted from taxg the etre surplus for several reasos. For example, re-electo motves o the part of poltcas may prevet the govermet from taxg all of the surplus from a strog costtuet group. Also, taxes are well oted for ther strog dscetve effects o actvtes that produce the taxed surplus, so polcymakers may be uwllg mpose taxes above certa rates. Addtoally, certa types of surplus may be hard to measure ad, therefore, more dffcult to tax. For example, t would be more dffcult to obta poltcal support for ad mplemet a tax o the ejoymet of evrometal servces tha a tax o corporate profts. Thus, corporatg some form of budget costrat cojucto wth lmted taxato should mprove the model, allowg for aalyss of these mportat effects. Exteded Model I ext develop a exteso of the basc model, corporatg a budget costrat ad lmted powers of taxato, as well as ecompassg multple propertes. The aalyss focuses o the regulator s decsos uder scearo 2. 5 There are several caveats to ths smplfcato. Govermets may also rase fuds by prtg moey or ssug debt. Sce ths artcle focuses o state govermets, whch lack the federal govermet s power to prt moey, I omt ths ssue. Smlarly, I omt debt ssuace for two reasos. Frst, my model s farly smple the tme dmeso, ad corporatg debt would complcate ths aspect cosderably. Secod, f oe assumes that debt must evetually be repad (.e. dsallowg bakruptcy ad debt-forgveess), the taxes rema the ultmate source for spedg. 16
Suppose that there are propertes uder the jursdcto of the govermet. Each property has ts ow prvate beeft, vestmet, ad socal cost as before ad the regulator makes a separate decso for each property, so the varables are dexed by property: p, I, s, r, C, 1,...,. I addto, the govermet has a budget costrat o the total compesato pad, r C t r s t 1 r p I s p, where r s a varable dexg regulato as before. r C s the compesato pad for property, wth C = p I uder a full-compesato rule ad scearo 2. t s < β s the tax rate o the socal beeft of (damage avoded by) regulato, ad t p < γ s the tax rate o owers prvate beefts. Parameters α, β, ad γ capture the govermet s ablty to rase tax reveue from varous sources. Parameter α s lump-sum fudg ot assocated wth ether prvate property values or the socal beefts of property. For a dvdual agecy, for example, ths amout represets the appropratos allocated to t from the state budget, whch do ot vary wth lad-use decsos. The coeffcet β < 1 shows the maxmum proporto of the socal beeft of laduse choces (.e. avoded socal cost) that the govermet s able to capture va taxato. The coeffcet γ < 1 shows the maxmum proporto of the prvate beeft of lad-use choces that the govermet s able to capture va taxato. For the govermet as a whole, these parameters reflect lmts o the govermet s taxato power due to poltcal costrats, the dffculty of measurg certa types of surplus, ad so forth. For dvdual ageces, these parameters could reflect creases budget allocatos based 17
o the performace of the agecy. Thus, varous combatos of α, β, ad γ would reflect dfferet assumptos about the ature of the govermet decso-makg etty. Frst Best Soluto As the basc model, a hypothetcal socal plaer s aggregate welfare-maxmzg decso provdes a bechmark for effcecy. The socal plaer attempts to maxmze expected aggregate welfare: prvate beeft less vestmet ad exteral socal cost, summed across all propertes. The plaer s ot costraed by a budget. PROPOSITION 6: Uder scearo 2, the socal plaer wll dsallow s developmet o parcel at tme 3 f ad oly f 1. p I Regulator s Decso As before, I assume that the regulator operates uder fscal lluso, gorg the prvate beefts of lad owers ad attemptg to maxmze the welfare of the rest of socety, et of taxes pad. The regulator s problem s expressed as 1 r s ts r s r Max, subject to the budget costrat, costrats o t s ad t p, ad the defto of C above. The form of the soluto to ths problem depeds o the ature of parameters α, β, ad γ. Therefore, I wll aalyze two represetatve cases of terest. Case 1 No Edogeous Tax Assume α > 0 ad β = γ = 0. These parameters descrbe the case of a partcular regulatory etty wth the larger govermet, oe that has o depedet taxato authorty ad reles o a fxed allocato for ts fudg. From ths fxed allocato, the agecy must pay ay compesato requred by ts regulatory decsos. The regulator s problem s to 18
choose a combato of parcels to regulate that wll avod the most socal harm whle keepg total compesato wth the fxed budget. The problem ca be expressed as Max 1 r s, subject to r p I. r Ths problem s aalogous to a beeft-cost problem wth a fxed budget for cost. The soluto ca be approxmated by frst rakg the parcels by a crtero such as the rato of socal beeft to requred compesato, p s I. The the regulator would choose to regulate the parcels from the hghest rato to the lowest, utl the budget s exhausted. From Proposto 6, effcet regulato requres that exactly those parcels for s whch 1 p I would be regulated. To the regulator, however, the bechmark rato of 1 s of lttle mportace. Istead, the sze of α plays the crtcal role. The regulator wll regulate parcels utl ts budget, α, s exhausted. Although t wll ted to regulate parcels wth hgh ratos ad allow developmet o those wth low ratos, there s o guaratee that the rato of 1 wll mark the boudary betwee the two groups of parcels. For relatvely large values of α, overregulato wll occur; that s, parcels wth a rato of less tha 1 wll be regulated because the regulator has more fudg tha s ecessary to compesate oly the parcels that would be effcetly regulated. For small values of α, uder-regulato wll occur; parcels wth a rato of greater tha 1 wll ot be regulated, because the tght budget s ot suffcet to compesate all of the parcels that would be effcetly regulated. 19
Cosder a specfc example. Suppose there are = 5 parcels wth the realzed values show table 1. If α = 10, the regulator wll regulate parcels 1 ad 2, because they gve the most socal cost avodace that s affordable wth a budget of 10. However, effcecy would dctate that parcel 3 also be regulated, sce ts rato of 1.25 s hgher tha 1. I ths case, therefore, uder-regulato occurs. If α = 20, the the regulator wll choose to regulate parcels 1-4, sce the total compesato requred, 20, s affordable uder the budget costrat, whle regulatg parcel 5 addto would ot be affordable. I ths case, however, overregulato occurs. Parcel 4 has a rato of less tha 1, ad effcecy requres that t ot be regulated. Parcel Prvate Beeft Ivestmet Compesato Socal Cost Rato () (p ) (I ) (C = p I ) (s ) s p I 1 5 2 3 6 2.0 2 10 5 5 8 1.6 3 7 3 4 5 1.25 4 10 2 8 6 0.75 5 5 1 4 2 0.5 Table 1 Example wth o edogeous tax Case 2 No Exogeous Fudg Suppose that α = 0, β > 0, ad γ > 0. Ths case could represet the stuato of the govermet as a whole, whch has taxed the maxmum amout possble from sources urelated to the use of the lad questo ad already allocated that moey to other 20
spedg. Thus, the govermet must rase ay fuds ecessary for compesato by taxg the prvate or socal beefts arsg from those parcels. The regulator s problem s expressed as Max 1 r s ts r s r p I t r s t r p I s p r 1., subject to the budget costrat PROPOSITION 7: Uder case 2 of the exteded model, the regulator wll regulate property f ad oly f p s I 1. Choosg to regulate aother parcel uder a bdg budget costrat mples that the regulator must rase taxes (t s ) o the rest of socety ot oly to pay the compesato o that parcel but also to replace the lost reveue from taxg that parcel s prvate beeft. s Recall that the effcet decso rule s to regulate f ad oly f 1. Uder- p I regulato occurs because 1 + γ > 1; the regulator wll ot regulate some parcels that would be regulated uder a effcet rule, those wth a rato betwee 1 ad 1 + γ. Cosder aga the umercal example, preseted table 2 wth addtoal formato o reveue from a tax o et prvate beefts. Assume t p = γ = 0.5. 6 The et prvate-beeft tax o some parcels ca be used to pay compesato to regulate others, eve wthout a tax o the socal beeft of the regulated parcels. Parcels 1-2 ca be regulated wth the compesato covered by the taxes leved o parcels 3-5, whch are allowed to be developed. Tax reveues o parcels 3-5 total 8, whch s equal to the total 6 As show the proof of Proposto 7 (see appedx), the regulator wll tax prvate beefts at the maxmum rate, uder the fscal lluso assumpto. 21
compesato requred for regulatg parcels 1-2. The regulator cosders whether to regulate parcel 3 also. Chagg that parcel from developmet to regulato decreases tax reveue by 2 ad creases compesato due by 4, for a total budget mpact of 6. Ths ca be accommodated oly by rasg taxes o the socal beeft. Sce the cremetal socal beeft s oly 5, the regulator wll ot regulate that parcel, eve though the socal value outweghs the et prvate value. Ths tradeoff s captured coveetly by the decso rule: regulate f ad oly f p s I 1. Sce parcel 3 s rato s 1.25, whch s less tha 1 + γ = 1.5, parcel 3 (as well as parcels 4 ad 5) wll ot be regulated. Sce parcel 3 would be regulated uder a effcet rule (1.25 > 1), uder-regulato occurs. Parcel Prvate Ivestmet Compesato Prvate Socal Rato () Beeft (p ) (I ) (C = p I ) Tax γ(p I ) Cost (s ) s p I 1 5 2 3 1.5 6 2.0 2 10 5 5 2.5 8 1.6 3 7 3 4 2.0 5 1.25 4 10 2 8 4.0 6 0.9 5 5 1 4 2.0 2 0.5 Table 2 Example wth o exogeous fudg Wthout restrctos o the values of β ad γ, a full-compesato rule wll geerally lead to uder-regulato. However, f the govermet ca be precluded from taxg the prvate beefts of developmet, γ = 0, ad t s able to tax a hgh proporto of 22
the socal beefts of regulato, 7 the effcecy ca be acheved. Ufortuately, such a proposto s lkely easer sad tha doe. Lmtg the tax o et prvate value may be possble, although t could mea elmatg the property taxes that are a mportat part of may state ad local tax structures. Empowerg the govermet to tax the socal beefts of regulato would lkely prove extremely dffcult. The poltcal opposto to such a tax could be tremedous, ad mplemetg a method for measurg such beefts ad apportog the tax amog taxpayers s fraught wth perl. Coclusos The prevous lterature o compesato for regulatory takgs s early uamous fdg that a full-compesato rule, requrg ucodtoal compesato equal to the market value lost to regulato, leads to effcet overvestmet o the part of the lad ower. Despte ths result, several states have cosdered or eacted legslato mplemetg such a rule. Ths behavor rases the questo whether such legslato s merely servg specal terests at the cost of effcecy or whether t ca be justfed as effcet uder models dfferet from those the prevous lterature. A smple model shows that, uder a alteratve sequece of evets, a fullcompesato rule ca be effcet. The key elemet of ths sequece s that owers are barred from vestmet utl the regulato questo s resolved. Ths pre-emptve regulato duces effcet choces by the lad ower. The full-compesato rule smlarly duces effcet regulatory choces, eve the face of fscal lluso, sce the 7 The ablty to tax the etre socal beeft, β = 1, cojucto wth γ = 0 would be suffcet, although ot ecessary, to acheve effcecy. The ecessary codto for effcecy s a β hgh eough, but possbly less tha 1, that settg t s = β just causes the budget costrat to bd. 23
ower s prvate beeft s teralzed. Thus, a full-compesato rule may be effcet uder some crcumstaces. If a state s commtted to usg such a rule, perhaps because of ts smplcty, these results suggest that mplemetg a pre-emptve restrcto that may be waved oce the relatve prvate ad socal values ca be compared would be complemetary to a full-compesato rule, creasg overall effcecy. Next, the model (usg the alteratve sequece) s exteded to corporate a budget costrat ad lmted taxato of the prvate ad socal values. The ature of the budget costrat ad taxato powers greatly affects the effcecy of a fullcompesato rule. If a dvdual agecy has a relatvely small budget ad o taxato authorty, the t wll be forced to prortze ts regulatory choces. Ths outcome has the desrable property that the regulatos eforced wll those cotrbutg most to aggregate welfare. However, f the budget s ot lked some way to the aggregate welfare geerated by the regulato, the ether over- or uder-regulato could occur. The govermet as a whole ca be thought of as rasg reveue from the taxato of the prvate ad socal beefts of lad use order to face ay ecessary compesato. Uder realstc assumptos about these taxato powers, however, such a desg wll lead to uder-regulato. Eve uder the favorable evromet evsoed scearo 2, a full-compesato requremet wll lead to effcecy whe cosderatos of budgets ad taxato are cosdered. If a state remas commtted to such a rule, the the results of the exteded model may provde some gudace for moderatg the effcecy of the rule. 24
Appedx Proof of Proposto 1 At tme 2, the vestmet level s gve. If o vestmet was made, the oly the udeveloped lad use s possble, for a aggregate welfare of 0. If I was vested to permt developmet, the the plaer must assess whether to proceed wth such developmet. Developmet wll yeld aggregate welfare of p s I, whle leavg the lad udeveloped yelds aggregate welfare of I. Therefore, the socal plaer wll maxmze aggregate welfare by proceedg wth developmet (f possble) at tme 2 f ad oly f p s > 0. At tme 1, the plaer must choose whether to vest I for potetal developmet, kowg the rule she wll follow at tme 2. She wll proceed wll developmet at tme 2 oly f p s > 0, so vestg I yelds a expected aggregate welfare of g s f p p sdpds I s, whle ot vestg yelds a expected aggregate welfare of 0. Thus, the plaer wll vest at tme 1 f ad oly f gs f p p sdpds I Ths decso rule maxmzes expected aggregate welfare by balacg the expected prvate beefts from developmet excess of socal costs agast the vestmet cost requred to allow the developmet opto. s. Proof of Proposto 2 The regulator s problem s Max 1 rs rc r. Uder the assumpto of fscal lluso, the regulator chooses r to maxmze the aggregate socal beeft excludg the property 25
ower s prvate beeft, less ay compesato requred. Uder the o-compesato rule, C = 0, so she must choose r to maxmze (1 r)s. Thus, she wll regulate (r = 1) at tme 2 wheever s > 0, wthout regard for the value of p. Uder a full-compesato rule, C = p, so the regulator must choose r to maxmze (1 r)s rp. Thus, she wll regulate (r = 1) at tme 2 f ad oly f s > p, whch s equvalet to the socal plaer s rule. Proof of Proposto 3 The ower s problem s de 1 r Max p rc I. Uder a o-compesato rule, C = 0, d so the prvate ower maxmzes de[(1 r)p I] kowg the regulator wll choose r = 1 wheever s > 0. Sce s > 0 s always true by tal assumpto, the ower maxmzes d( I). Thus, he wll choose d = 0 all cases. Uder the full-compesato rule, C = p, so the ower chooses d to maxmze de[p I], sce he receves p I f he chooses to develop the lad, regardless of the subsequet regulato decso. Thus, he wll choose developmet (d = 1) f ad oly f p I 0, whch s true by tal assumpto. Thus, he wll always vest for developmet. Proof of Proposto 4 Developmet wll yeld aggregate welfare of p s I, whle leavg the lad udeveloped yelds aggregate welfare of 0. Therefore, the socal plaer wll proceed wth developmet f ad oly f p I s > 0. 26
Proof of Proposto 5 The regulator s decso problem s aga expressed as 1 rs rc Max, where r = 1 whe the regulator chooses to regulate (.e. prevet developmet) ad r = 0 otherwse. I scearo 2, a full-compesato rule takes the form C = p I, sce the loss due to regulato s adjusted for avodg the vestmet cost. Uder ths rule, the regulator chooses r to maxmze (1 r)s r(p I). Thus, she wll allow developmet at tme 3 f ad oly f p I > s. r Proof of Proposto 6 1 The plaer s problem s expressed as Max 1 r s r p I r. The plaer s ot subject to the budget costrat. The dervatve of the objectve fucto wth respect to r s s p + I. Thus, the plaer wll regulate (r = 1) f ad oly f s s p + I > 0, whch s equvalet to 1. p I Proof of Proposto 7 Max 1 r s t The regulator s problem s expressed as s budget costrat p I t r s t r p I s r p r s, subject to the r 1. Sce t p does ot affect the objectve fucto, the regulator wll tax prvate beefts of lad use at the maxmum rate, t p = γ. She wll also lower t s utl the budget costrat bds wth equalty, p I t r s r p I s r 1. Substtuto of the bdg budget 27
costrat to the objectve fucto smplfes the problem to 1 r s r p I 1 r p I Max. r The dervatve of the objectve fucto wth respect to r s s (p I ) γ(p I ). So the regulator wll choose r = 1 f ad oly f s (p I ) γ(p I ) > 0, ad r = 0 otherwse. Ths codto s equvalet to the decso rule to regulate (r = 1) f ad oly f p s I 1. 28
Refereces Blume, L., D. Rubfeld, ad P. Shapro, 1984. The Takg of Lad: Whe Should Compesato Be Pad? Quarterly Joural of Ecoomcs 99: 71-92. Buchaa, J.M., 1962. Poltcs, Polcy, ad the Pgova Margs, Ecoomca 29: 17-28. Hermal, B., 1995. A Ecoomc Aalyss of Takgs, Joural of Law, Ecoomcs, ad Orgazato 11: 518-537. Ies, R., 1997. Takgs, Compesato, ad Equal Treatmet for Owers of Developed ad Udeveloped Property, Joural of Law & Ecoomcs 40: 403-432. Ies, R., S. Polasky, ad J. Tschrhart, 1998. Takgs, Compesato, ad Edagered Speces Protecto o Prvate Lads, Joural of Ecoomc Perspectves 12: 35-52. Mart, S.A. ad K. Shrver, Documetg the Impact of Measure 37: Selected Case Studes, prepared Ja. 2006. http://www.pdx.edu/ms/m37.html. Mcel, T. ad K. Segerso. Compesato for Regulatory Takgs: A Ecoomc Aalyss wth Applcatos. Greewch CT: JAI Press. 1996. Mcel, T. ad K. Segerso, 1994. Regulatory Takgs: Whe Should Compesato Be Pad? Joural of Legal Studes 23: 749-776. 29