"Business, Sauder School of"@en . "DSpace"@en . "UBCV"@en . "Carruthers, Cameron M."@en . "2008-09-17T17:16:05Z"@en . "1993"@en . "Master of Science in Business - MScB"@en . "University of British Columbia"@en . "Urban areas such as the City of Vancouver offer employment opportunities and living amenities that are attractive to other Canadians and immigrants. The result is a rate of growth that exceeds the national average. It is the responsibility of the elected representatives to implement public policy that will provide housing for these new residents. As the amount of available land diminishes, increased density plays a more significant role.\r\nThe purpose of this paper is to determine how, if any, increased density impacts the value of adjacent properties. It has been argued by individuals and local area interest groups that property values will decline if increased density residential development is allowed to occur in their neighborhoods. One such neighborhood, the Kitsilano neighborhood in Vancouver's west side, is undergoing a rezoning from light industrial to higher density residential relative to the adjacent neighboring properties. Using a statistical procedure known as hedonic modelling, the results indicate that the rezoning has had no impact on property values. This could result from the fact that the rezoning indeed has no effect on property values as is evidenced by a great deal of the literature on the subject. However, the model may be mis-specified and these results may be interpreted in a number of other ways. For instance, the rezoning studied in this paper has two components. The area is experiencing a land-use change from light industrial to residential which theoretically ought to increase property values in a predominantly residential neighborhood. The second component involves the proposed increased density relative to the existing neighborhood with an anticipated negative effect. The results of these two forces may be cancelling each other out in the empirical model."@en . "https://circle.library.ubc.ca/rest/handle/2429/2148?expand=metadata"@en . "3297963 bytes"@en . "application/pdf"@en . "THE IMPACTS OF REZONING ON ADJACENT PROPERTY VALUESbyCAMERON MARTIN CARRUTHERSB. Comm., The University of Alberta, 1990A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF SCIENCE (BUSINESS ADMINISTRATION)inTHE FACULTY OF GRADUATE STUDIES(Department of Commerce)We accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAOctober 1993\u00C2\u00A9 Cameron Martin Carruthers, 1993In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature Department of atiMEize.EThe University of British ColumbiaVancouver, CanadaDate 4/-0,5ek. /1'09193DE-6 (2/88)AbstractUrban areas such as the City of Vancouver offer employment opportunities and livingamenities that are attractive to other Canadians and immigrants. The result is a rate of growththat exceeds the national average. It is the responsibility of the elected representatives toimplement public policy that will provide housing for these new residents. As the amount ofavailable land diminishes, increased density plays a more significant role.The purpose of this paper is to determine how, if any, increased density impacts thevalue of adjacent properties. It has been argued by individuals and local area interest groupsthat property values will decline if increased density residential development is allowed to occurin their neighborhoods. One such neighborhood, the Kitsilano neighborhood in Vancouver'swest side, is undergoing a rezoning from light industrial to higher density residential relative tothe adjacent neighboring properties.Using a statistical procedure known as hedonic modelling, the results indicate that therezoning has had no impact on property values. This could result from the fact that the rezoningindeed has no effect on property values as is evidenced by a great deal of the literature on thesubject. However, the model may be mis-specified and these results may be interpreted in anumber of other ways. For instance, the rezoning studied in this paper has two components.The area is experiencing a land-use change from light industrial to residential which theoreticallyought to increase property values in a predominantly residential neighborhood. The secondcomponent involves the proposed increased density relative to the existing neighborhood withan anticipated negative effect. The results of these two forces may be cancelling each other outin the empirical model.Table of ContentsAbstract ^ iiTable of Contents ^ iiiList of Tables ivList of Figures v^Chapter 1^Introduction ^ 1Chapter 2^Review of the Literature ^ 72.1^Zoning and Externalities - An Introduction ^ 72.2 The Economic Theory of Zoning 112.3^Empirical Research on Zoning and Property Valuation ^ 192.3.1 Zoning and Property Values in a Local Context 222.3.2 Additional Research on the Externality Argument ^ 272.3.3 Higher Density Housing and Property Values 30Chapter 3^The Study ^ 343.1^Kitsilano: A Brief History ^ 353.2^Kitsilano Today: The Arbutus Industrial Area ^ 383.3 Methodology ^ 413.4 The Data 433.5^The Study Area ^ 453.6 The Control Area 483.7^Discussion of the Study Area and the Control Area ^ 513.8^Regression Results ^ 52Chapter 4^Conclusion 58References ^ 62111List of TablesTable 2.1 ^ 12Table 3.1 44Table 3.2 ^ 46Table 3.3 50Table 3.4 ^ 55Table 3.5 57ivList of FiguresFigure 2.1Figure 2.2Figure 3.1Figure 3.2Figure 3.3VChapter 1 IntroductionIt is a common occurrence that when redevelopment or rezoning applications comeunder scrutiny of the general public, there is an outcry from local interest groups. Evidenceof this is abundant and can be found by reading the local newspaper or minutes of localcouncil meetings. A proposal to subdivide a single lot to provide two single familydwellings while still within the existing minimum lot requirements has led to such chargesas \"plummeting property values\" and \"cheapened neighborhood appearance\". By makingpublic their intention to sell part of their land holdings to the British Columbia HousingCorporation for 21 affordable townhomes, a local church has met with neighborhoodopposition who fear that it would lower property values in the area?. The City ofVancouver's plan to rezone the Arbutus Industrial Area to provide affordable housing withinthe central area has met with resistance by residents in the surrounding area who areconcerned about the destruction of the \"west-side\" style of life3.Rezoning reports such as the ones referred to above typically use such terms as'affordable housing' and 'increased density' interchangeably. With the land component'The Vancouver Sun, Tuesday July 27, 1993, page B3. \"70 Oxford Heights Residents ProtestProposed Subdivision\". They fear it would drastically change the character of the area and causeproperty values to plummet. They ask for local government protection from unscrupulousdevelopers.'The Vancouver Sun, Monday August 24, 1992, page Bl. \"Residents Picket Church to tryto Stop Housing Project for the Poor\". The church argues there is a real need for affordablehousing yet neighbors fear the increased density will destroy the visual appeal of the neighborhoodand will take away from the value of their homes.'The Vancouver Sun, Tuesday January 13, 1992, page B3. \"Arbutus-Area Residents Fightfor Local Lifestyle\". The neighborhood opposition group says that rezoning from light industrialto higher density residential relative to the low density residential neighborhood is \"illogical andharmful\". They indicate there is a lot of evidence to show that high density cities don't work.1currently playing such a significant role in many areas of Vancouve, an increase in thebuilding density effectively reduces the per unit cost of land.As noted, the property value argument is not the only one brought into the politicalarena by those opposed to neighborhood change. Other concerns such as increased noise andtraffic, increased crime, decreased levels of safety and security, and a general loss ofcommunity values have been raised to oppose such neighborhood changes as increaseddensity. This collection of arguments encompass a common theme, namely a resistance tochange.People's attitudes are generally resistant to change (Worchel, 1988). Thisphenomenon can be observed in all aspects of peoples lives. Whether examining interactionpatterns between family members, cohort relationships in the workplace, or resident'sassociations with their neighbors a degree of consistency is apparent in the observedbehaviour patterns. However, people usually deny that they are unyielding to change. Anincrease in density in a particular neighborhood will bring a flurry of responses, many ofthem negative. Simply citing an unwillingness to change is considered an inadequateresponse, a response that has no justification. Cite plummeting property values though andsuddenly the response appears justified.People are not likely to become less resistant to change. As the population ages wefind ourselves less willing to cope with change. In his analysis of population and housingin Vancouver, David Baxter (1989) indicates that,'The Vancouver Sun, Thursday September 16, 1993, page El.2It is the demographic process of the aging of the post war baby boominto the 35 to 44 age group (1986-1990) and then into the 45 to 54 agegroup (1991-2006) that will determine the characteristics of changesin housing demand in metropolitan Vancouver. . . . This will meanchange in our communities, such has happened before in othermetropolitan areas as they have grown, with increased density ofdevelopment in highly accessible locations and new development insuburban locations (p. 79).Anyone working in a profession that advocates change must therefore tolerate and be ableto deal with opposition. For instance, urban planners in a growing area such as the LowerMainland are going to experience a great deal of pressure from neighborhood groups tomaintain the status quo. The Not In My BackYard' syndrome is a common response toincreased density and affordable housings.However, change is inevitable. People have an impeccable ability to proliferate. Theglobal population is reaching the five billion mark and is continuing to grow. The populationhas to be housed. People migrate to areas that offer jobs. There appears to be an increasingdesire to move to areas that offer amenities for a desired lifestyle (Gappert and Knight,1982). Change will be faster and more sudden in the areas that offer jobs and amenities.The Lower Mainland is currently such a place. The Vancouver Planning Department hasrealized that this change is inevitable and in June of 1986 produced a document investigatingthe reaction of neighbors to increased density in their neighborhoods (City of Vancouver,1986).The above discussion brings us to a typical string of events. People are going to'Vancouver Sun, Friday December 11, 1992, p.B1. \"Affordable Housing a Right, ResidentsSay\". In a recent survey, two thirds of Greater Vancouver residents say affordable housing is a'Canadian right'. Forty percent blame government while 21 percent blame developers for lackof affordable housing. However, fifty-five percent do not want to see increased density in theirneighborhood.3move into specific urban areas that offer jobs and desirable amenities. This appears to beinevitable. Existing density patterns will be unable to accommodate an increased population.Planners and civic officials will seek out opportunities to increase density in order to provideaffordable housing. Current residents will oppose proposed changes to their neighborhoods.They will attempt to thwart density change using such arguments as declining propertyvalues, decreased amenities or increased costs. This conservative stance appears to be fact.If private property values are indeed affected by such an external action as rezoningan adjacent property, there is an argument that there ought to be some form ofcompensation. This appears to be the basis for the Uthwatt Committee in Britain in 1943when determining how to compensate private landholders for post-war reclamation projects(Young, 1943). In their report they consider both compensation and betterment, the formerwhen the external action negatively affects property and the latter when the external actionpositively affects property. External actions therefore create both winners and losers. Itseems rather obvious that if neither winner nor loser were responsible for the external action,the winner ought to forfeit his or her betterment and the loser be compensated. Yet not allexternal actions produce a tidy list of winners and losers. For instance, when rezoning aparcel of land that is surrounded by a variety of land uses, the ramifications will be felt bothnear and possibly far from the site. The total assessed values of the area, however defined,may or may not remain the same.The purpose of this paper is not to alter any of the above facts nor is it to determinean assured compensation scheme. The purpose of this paper is to address the merits of oneparticular argument. This argument can be stated in the following manner:4\"A rezoning to a higher density use has a negative impact on adjacent property values\".The intention here is to effect a better understanding of this argument from a pool ofrelated literature. Chapter two reviews the economic literature on externalities and zoningand how they affect property values. Although the numerous studies have come to differentconclusions, the dominant result is that there are no apparent significant impacts on propertyvalues. In addition, chapter two determines the important elements that are to be employedin a statistical analysis of a rezoning on adjacent property values. The studies that arediscussed in Chapter two consider a variety of external actions and how they affect propertyvalues.Chapter three incorporates the elements gleaned from chapter two and uses them toanalyze the current rezoning issue facing the residents of the Kitsilano neighborhood in theCity of Vancouver. A section of this neighborhood, the Arbutus Industrial Area (AIA), isa 10.4 Hectare (25.8 acre) block of currently zoned light industrial land situated south ofBroadway and west of Arbutus Street. The need for affordable housing in the region hasprompted Vancouver City Council to identify areas for infill high density housing. The AIAis undergoing a land use transition which has local area residents in opposition with theplanning department and the land owners of the site. Statistical analysis on the determinantsof property values and interpretation of the results form the basis of the chapter. Theresults, like so many of the studies outlined in Chaper two, indicate no statisticallysignificant impact on property values due to the anticipated rezoning. This could stem fromthe result that the rezoning has not been approved and is still facing public debate.5Finally, chapter four concludes the study by reiterating the salient outcomes as wellas cautioning the reader on interpretation of the results.6Chapter 2^Review of the Literature2.1 Zoning and Externalities - An IntroductionThis paper is concerned with the impacts of density on the spatial separation of landuses. The discussion that follows will direct its efforts towards this particular aspect ofzoning. It must be noted that zoning is seldom done in isolation from such other forces asdevelopment restrictions, setbacks, and height restrictions to mention only a few.Zoning, as the term implies, is the segregation of land into a variety of uses. Broadlyspeaking, these uses are frequently divided into residential, commercial, industrial and publicspace. Each of these can further be subdivided as determined by the local governing body.For example, residential zoning in Vancouver can be seen as a combination of single family,duplex and/or conversion, and apartment.There typically have been three main arguments justifying zoning in an urban area(Mark and Goldberg, 1981, p.419):1. The existence of negative externalities in urban landmarkets and the corresponding need to control such externalitiesby ensuring that incompatible land uses do not exist in closeproximity with each other. Another variant of this externalityargument is that there is a need to preserve and protect propertyvalues from external effects and that exclusive use zones can dothis. (The externality argument).2. The haphazard development pattern resulting from marketforces leads to leapfrogging of development and sprawl. Suchinefficient patterns of development lead to high costs of servicingand leave undeveloped pockets of land. Zoning, it is argued, canprovide for more efficient development patterns. (The spatial7efficiency argument).3. Market forces also lead to speculative cycles and toneighborhood instabilities. By regulating land use and land usesuccession through zoning controls, a more orderly market can beachieved (according to zoning proponents) with preservation ofneighborhoods and orderly evolution of land uses over time. (Theintertemporal efficiency argument).From an economic standpoint, zoning is presented as a tool that reduces or eliminatesthe inefficiencies in the unregulated land market. These inefficiencies have been referredto in the literature as externalities, external economies, external diseconomies and spillovers.According to the MIT dictionary of modern economics, an externality is defined as the\"interdependence of utility and/or production functions...which has not been accounted forthrough trade\" (Pierce, 1986). The term 'neighborhood effect' implies that theseexternalities have a spatial component. For example, the noise generated from the operationof a night club would constitute a negative neighborhood effect on nearby residents while abeautiful park would constitute a positive neighborhood effect.Whether zoning has mitigated such neighborhood effects as higher density has beenthe subject of many empirical studies with differing results. Kain and Quigley (1970), Avrin(1977), Jud (1980), Peterson (1974) and Stull (1975) have all found some evidence thatexternalities impact property values. On the opposing side, Crecine, Davis and Jackson(1967), Reuter (1973), Grether and Mieszkowsld (1980) and Mark and Goldberg (1981,1986) have all found inconclusive evidence of such effects. The benefits of zoning appearmore pronounced when segregating seemingly incompatible land uses such as heavy industryand residential. On a more subtle level, segregating single family residential from duplexmay be regarded as questionable from an efficiency standpoint.8In a metropolitan area such as Vancouver where there has been moderate growth inrecent years (Baxter, 1989) and little room for suburban expansion, the concept of rezoningbecomes critical. Add to this the earlier cited evidence implying increased density in thecentral urban area is an expected result in metropolitan growth and rezoning becomesdemand driven.Rezoning is typically classified as up-zoning, down-zoning or different land usezoning. Although these terms have been used interchangeably, the term up-zoning will beused when there is an increase in the allowable density or when a less desired use is allowed.In the Arbutus Industrial Area rezoning, there has been what some consider down-zoning asresidential development is more desirable than light industrial. However, some consider therezoning an up-zoning issue as the density of the area is being increased relative to theadjacent residential neighborhood.There are both public and private costs associated with rezoning. There is little basisto refute the notion that increasing density imposes private gross costs on the existingresidents. An example of private costs is the increased traffic congestion as a result of theadded density. Less obvious and somewhat questionable with respect to private costs are theloss of views that a higher density structure would impose on the neighborhood and theincreased use of the public facilities such as schools and parks. An example of a public costof up-zoning would be the additional infrastructure requirements to meet the standards setby the local governing agency.Of course there are also benefits, both private and public, associated with up-zoning.Increased density may trigger the establishment of convenience stores which would reduce9travel time, and hence be considered a private benefit. Efficient use of public infrastructureis often achieved only after a given level of density (enrolment in schools). There is alsopublic benefit to housing residents in an affordable manner. Economists argue that thecombination of private costs and public costs should be compared to the combination ofprivate and public benefits when assessing public policies. However, a more commonoccurrence posits private costs against public benefits in the public arena.In addition to the economic and social implications of rezoning, there are politicalimplications as well. Council ought to act in the best interests of the population as a wholewhen approving public policy. However, council also wants to be voted for in the nextelection and will not want to dissuade potential supporters nor provoke dissenters. It is alsoat the political level that input from interest groups impacts the decision making process.For these reasons, the political implications with regards to rezoning are difficult to quantify.The above discussion encompasses the economic, social and political implications ofzoning with respect to public policy. The review of the literature that follows focuses on theeconomic effects of zoning and on various neighborhood effects, both of which are perceivedto have an effect on land values. A thorough understanding of this literature and effectiveinteraction between local government and the public may enable a more efficient outcomefrom the public policy forum.102.2 The Economic Theory of ZoningAs an impediment to the free market mechanism, zoning is an ideal candidate forstudy by economists. However, little work has been done in the area as is evident by thescarcity of literature on the theory of zoning relative to such urban issues as propertytaxation. Zoning as a regulation of land use is based on the fact that \"some forms of landuse in any circumstances, and most forms of land use in some circumstances, havebeneficial or harmful effects on neighboring properties\" (Bailey, 1969). Since the turn ofthe century\u00C2\u00B0 zoning has become an omnipresent phenomenon in urban North America. Ithas only been since 1974 with the work of such economists as William J. Stull, James Ohls,Richard Weisburg and Michelle White that zoning has become a focus of mathematicalmodelling.In their review of the theory of zoning, Pogodzinski and Sass (1990) have categorizedthe effects of zoning. Although neither mutually exclusive nor exhaustive, Table 2.1 coversthe important urban economic categories. This disaggregation of zoning impacts is tosimplify analysis and in no way implies one category is more important than another. Abrief discussion of the impacts of zoning on each category provides an indication of thecomplexities and inter-relatedness of studying zoning.Supply side and demand side effects typically form the basis of economic analysis.An example of a supply side effect would be the restriction of available land for a'Zoning and its derivatives originated in New York City in 1916. Refer to Goldberg andChinloy (1984) for a brief discussion on the evolution of property rights in North America.11TABLE 2.1CategorySupply Side EffectsDemand Side EffectsTiebout EffectsExternality EffectsEndogenous ZoningRent-seeking BehaviorImpactsSuppliers of Housing ServicesConsumption of Housing ServicesConsumer MobilityNon-conforming Uses / CongestionIntroduction of Zoning as a Land Use RestrictionSelf-interest use of Zoningparticular use. If, for example, large-lot zoning was enforced in a particular jurisdiction,then given demand levels, the cost of producing housing services would increase. Courant(1976) shows as a result, that housing prices rise and welfare falls. Henderson (1985) alsoconsiders large-lot zoning as a restriction on the supply side but extends the analysis byexamining this effect upon the demand for housing. Henderson shows that in the presenceof zoning, average total costs of a quality adjusted unit are greater.Consumer mobility is the focus of the Tiebout Hypothesis' and becomes integral ineconomic analysis when the assumption of equalized utility levels across jurisdictions isimposed. The ability to move intra or inter-jurisdictionally is appealing to residents in theevent of a favorable zoning restriction in another equally desirable neighborhood orcommunity. White (1975) attempts to determine the effects of zoning on the utility levelsof central city versus suburban dwellers. Although her assumptions have been criticized byacademics, her modelling provides insight into income discrepancies and the role zoningplays in maintaining this disparity.'The Tiebout Hypothesis states that the existence of numerous competing local governmentsprovides consumers of local services a large number of choices regarding residential and worklocation. This competition evokes local government efficiency if residents are mobile.12The ability to incorporate endogenous zoning into the economic model provides avaluable tool for evaluating zoning reform. Epple, Romer and Filimon (1988) develop amodel based upon Henderson's (1980) work on community development. They endogenizezoning in order to determine whether exclusionary zoning' occurs under various decisionmaking regimes. Additionally, they analyze the effectiveness of development fees9 as amethod of dealing with inefficient land use.As an extension of the endogeneity of zoning, wealth maximizing individuals willexpend resources influencing civic officials to pass zoning by-laws favorable to them. If thiswere the case, zoning as a socially beneficial regulation becomes questionable. This rent-seeking category is the focus of Fischel (1978) in his discussion of private property rights.Zoning is considered a transfer of property rights from the individual land owner to the civicauthority or, through market forces, to another land owner. In an attempt to minimize thecosts of maintaining neighborhood quality, the existing residents rely upon zoning to restrictthe owners of vacant land from 'over-development'. The result is that restrictive zoningcreates a lower market price of land than in the absence of zoning, as the existing residentsdo not realize the actual opportunity cost of land. This, in turn, leads to a smaller thanoptimal community size.The final category examines externality effects and will dominate the remainder ofthis section. As the focus of this paper is on the effects of a 'non-conforming' use on'Exclusionary zoning has been defined as \"zoning which results in land in some incorporatedcommunity being unoccupied, while higher cost land is developed\" (Epple, Romer and Filimon,1988, p.135).'Development fees, lot levies or subdivision exactments intend to defray the cost of certainpublic-sector capital outlays specific to the new development. For a more detailed description,refer to Bird and Slack (1983).13adjacent property values, the contributions set forth in this section are emphasized.Stull (1974) is one of the first economists to employ zoning into an urban economicmodel. He constructs a two sector model (manufacturing and residential) in an attempt showto how a \"change in the allocation of land between the two sectors affects the spatial patternof land rents, the size of the urban labor force, and the wage rate\" (Stull, 1974, p. 337).The spatial pattern of land rents are of interest in this discussion as decreased property valuesin the residential sector resulting from an increased size of the manufacturing sector wouldbe considered a negative externality.Stull's model is based upon a number of simplifying assumptions. The firstassumption is that of a linear city. There is a fixed central transportation node from whicha featureless line extends infinitely out in two directions. The only connection with theoutside world is from the central node. The second assumption is that all transactions occurat this central node. There are only two markets operating in this city, a labor market anda land market. Residents must commute to this central node thus incurring commuting costswhich are a function of the distance from the central node. The third assumption is thatthere are only two uses for land, as a location for residential activities and as a location formanufacturing activities. Fourthly, there is a profit maximizing developer who is in controlof the spatial structure of the city. Assumptions one and three are for simplification whileall the assumptions are typical of the literature. As can be seen in Figure 2.1, the developerallocates the portion from t ti: to manufacturing activities and the two portions, tb' tot: and t: to t\u00E2\u0080\u009E ' to residential activities. This spatial allocation embodies the tenet thatresidential and manufacturing activities do not mix. All land to the left of tb' and to the14right of remain agricultural.The positions of tba and t: are dependent upon the profit maximizing developeroptimizing the spatial relationship between the manufacturing and residential sector.Residents must commute and work in the manufacturing sector, yet presumably do not desireto live in close proximity to the manufacturing sector. This is where the neighborhoodexternality effect becomes apparent.t:^C^ta*Figure 2.1In deriving the model, all residents are assumed to have identical utility functions asfollows:u = u(xi , x2, ...^ (2.1)where:^x= the quantity of ordinary commodity i purchased,q = the quantity of land rented,t = the distance from the consumer's residence to the utaicentre,t. = the distance from the nearest zoning boundary to theurban centre.The last component in the utility function introduces the neighborhood externality effect.15The first order partial derivatives are assumed to be:uxi^>0^(i=1,2,...,n)Uq^>0ut^<0ut-t\u00E2\u0080\u00A2 >0The last partial derivative implies that consumers prefer living farther away from a zoningchange boundary. This consumer preference and the intensity of its impact on propertyvalues forms the basis of this study.Each resident is also faced with a budget constraint of the following form:y = pixi + mq + h(t)^ (2.2)where:Pi==h =the annual urban money wage rate,the implicit or hedonici\u00C2\u00B0 price of the residentialproperty attributes,the annual land rent payment per unit of land,the transportation cost function.Given the three functional restrictions that 1) every household must always be within itsbudget constraint, 2) the quantity of land purchased is fixed in advance and constant at alllocations, and 3) the utility level is fixed, together with the assumption of consumerrationality; the land rent gradient for the residential sector can be derived in the followingform:'For the foundational literature on hedonic price theory, see Rosen (1974).16m = m(t; y , t* )^ (2.3)Similar analysis can derive the land rent gradient for the manufacturing sector as follows':M = M(t; y)^ (2.4)Figure 2 provides a graphical representation of these rent gradients. The slope of the rentgradient for the residential sector can be derived through use of the envelope theorem andis expressed as follows:a rn/at = [ 1/q*][ -h'(t)] +^14, +^(2.5)As is evident in Figure 2.2, the sign of this expression is dependent upon the externalityeffect. If there is substantial opposition to residing near the manufacturing sector, then theresidential rent gradient may behave as depicted in Figure 2.2. However, the depiction inFigure 2.2 is not consistent with equilibrium in the land market. If other economic factorssuch as commuting costs play a more salient role in the consumer's utility function, a moretraditional downward sloping continuous rent gradient may be observed (refer to dashedline).\"In his paper, Stull (1974) leads the reader through the mathematics of deriving both of theseland rent gradients.17Figure 2.2The purpose of this section has been to validate the legitimacy of zoning as acandidate for economic research. The theory argues that there are many economic categoriesthat have been incorporated into theoretical modelling. One of these categories is the effectsof externalities. The important result to draw from the theory on externalities and zoningis that the distance between two different land uses may have an impact on their respectiveproperty values. This impact is responsive to the degree of incompatibility of land uses andthe intensity of other economic factors.182.3 Empirical Research on Zoning and Property ValuationAlthough neighborhood effects have been the topic of academic research for over fiftyyears (Haig, 1926), the first attempt to quantify these effects statistically is in researchcarried out by Crecine, Davis and Jackson (1967). They use the available data from selectcensus tracts in the city of Pittsburgh from 1956 to 1963 to determine the nature and extentof neighborhood externalities. In order to view neighborhood effects, given that the urbanproperty markets are competitive, Crecine, Davis and Jackson assume that buyers havesimilar tastes. If individual tastes were diverse, an efficient and competitive market wouldimply that what some deemed a diseconomy, others would consider an economy and priceswould reflect neither the diseconomy nor the economy. In the context of the ArbutusIndustrial Area rezoning, those who deem the rezoning a diseconomy will move from theneighborhood and those who do not view the rezoning as a diseconomy will move into theneighborhood.In establishing their model, they divide their study into 'insular' neighborhoods byreferring to the land use block in which the sale took place. Thus, a sale price would beaffected by an externality within the determined land use block, but would not be affectedby an externality that was across the street from the subject property. This implies that thereare no spillovers, an assumption that has been questioned by subsequent research (Strange,1992). The model formulation adopted by Crecine, Davis and Jackson follows one definedby Brigham (1965). The value of a site is a function of access, amenities, topography,present and future use, and other historical factors. Mathematically this can be defined as19follows:Ai^,^ )^ (2.6)where Vi = the value of the 1th site,= the eh sites' accessibility to economic activities,Ai = the 'amenities' (including neighborhood effects) of the eh site,= the topography of the eh site,= the present and future use of the eh site,= historical factors affecting the utilization of the eh site,This general framework, known as hedonic modelling, forms the basis of much of theresearch in property valuation. This model must be modified to take into account theavailable data. In the Crecine, Davis and Jackson paper, the dependent variable used is aprice per square foot of the site. This is regressed on a vector of zoning externality variablesdefined as the percentage of total land area in the block associated with a particular land use.Additionally, a vector of non-zoning externalities and a monthly index are also included asindependent variables. In essence, their model stipulates that the price per square foot ofsingle family dwelling site is a linear combination of time and externalities. Their resultsimply that the property market is independent of externalities. It is cautioned that this resultmay be partially due to the specification of the model. The authors imply that theneighborhood effects could extend beyond the defined block (spillover), or could beconsidered \"next door\" phenomenon (p. 93). Thus, the arbitrary division of the city intoinsular neighborhoods fails to accommodate the spatial dimension of the externality. Anotherexplanation for the results is that prospective purchasers have heterogeneous tastes. Anexternality may be deemed undesirable by some, yet be considered desirable by others.20Those that do not find a particular land use undesirable will be indifferent to locating nearit and the price paid for their property will thus not reflect the externality.The results espoused by Crecine, Davis and Jackson imply that zoning, in its currentform, is perhaps not the most appropriate vehicle to handle neighborhood externalities. Thisresult has stimulated subsequent research into the effectiveness of current zoning practices.Grether and Mieszkowsld (1980) consider the weaknesses in the Crecine et. al. study andattempt to overcome them. They criticize using only lot size as a control for individualproperties and propose using a vector of variables representing various structural, lot and agecharacteristics. They also attempt to overcome the 'insular' neighborhood problem byutilizing a continuous distance measure from each property sold to the particularneighborhood externality for each of 16 experiments. Their results are consistent withCrecine et al. in that there are no statistically significant effects on housing prices due to theproximity of a variety of minor commercial developments and low density apartments.However, they have a problem with sample size, with the number of observations in someof their experiments being less than 50.In addition to finding similar results to the Crecine et al. study, Grether andMieszkowski (1980) provide insight into model specification. Two models are discussed intheir paper. Their first model \"allows for different appreciation rates for structural andnonstructural characteristics, the cost being that a nonlinearity is introduced\" (p. 7). Theirsecond model is a semilogarithmic specification, and is used in the analysis partially due toits simplicity. In an earlier paper, Grether and Mieszkowsld (1974) provide an extensivediscussion on the hedonic formulation they have incorporated in their 1980 study. It is a21refined version of the model described by equation 3.1 and can be formulated as shown:log = Sia + Lifl + Islzy + ei^(2.7)where:^log vi is the natural logarithm of actual sales price,Si is a vector of structural characteristics,Li is a vector of lot characteristics, andNi is a vector of neighborhood characteristics.\u00C2\u00AB,13,-y, and ci are the coefficients and error term.Whether to use a semilogarithmic specification, a nonlinear specification or some othervariant as the correct specification in these studies on property values has been a topic ofdebate. The semilogarithmic form has economic appeal as it implies that the independentvariables provide a constant percentage change to the value of the house. For this reason,the semilogarithmic specification has been used extensively in the research and will beconsidered in this study.2.3.1 Zoning and Property Values in a Local ContextBoth the impacts of zoning and rezoning on property values have been analyzed inthe Vancouver area. While both papers in this section focus on the impacts of zoning on theactual property, they shed insight on how to look for neighborhood effects. In their firstpaper on the topic, Mark and Goldberg (1981) consider the impacts of rezoning and actualproperty values. In addition, they want to assess the ability of zoning to \"mitigateexternalities\" (p. 418). Their paper considers three separate methodologies and three studies22performed in the Vancouver metropolitan area. The first study was done by Tunnicliffe(1975) as an M.Sc. thesis and incorporated all rezonings in the City of Vancouver for aperiod extending from 1966 to 1972. He looks at property values before and after arezoning and compares the change in value to a comparable change in property values in acontrol group. Although a departure from the standard statistical analysis discussed above,Tunnicliffe finds two interesting results. The first is that 'up-zoning' does not necessarilylead to higher property values. This seems to contradict the view that increased densityallows increased revenue generation per unit area of land, which should lead to increasedproperty values. The second result is that there does not seem to be consistent increasesfrom similar rezonings. Both of these results imply that it would be difficult to tax theaffected property owners in order to redistribute the benefits gained from the rezoning. Oneflaw in this study that has been addressed in subsequent research is that assessed values areused rather than actual sales data.The next study of the Vancouver region considered the rezoning of the Kerrisdaleneighborhood and incorporated data from 1956 to 1966 12 . This area had been exclusivelysingle family low density, and in 1961 a portion of the area was upzoned to multi-familymedium-density (10-12 story apartment buildings). The study consists of breaking theaffected area into seven concentric half block bands in addition to four independent areas ascontrol. Actual sales data on 979 properties that sold before and after the rezoning are used.It was found that properties greater that 5000 square feet show a significant increase in value'For a complete discussion regarding this study, the reader is referred to: Goldberg, Michaeland P. Horwood, Zoning,: Its Costs and Relevance for the 1980's, Vancouver: The FraserInstitute, 1980).23due to the rezoning than smaller properties. This makes intuitive sense as the larger lotswould be viewed by developers as ideal sites for redevelopment. They also find nosignificant neighborhood effects imposed on the surrounding single family dwellings. Again,this finding is consistent with the Crecine, et.al . study. A third result implies that there islittle evidence that speculative activity increased due to the rezoning.The final study area examined in the Mark and Goldberg (1981) paper is the city ofNew Westminister. This study considers the impacts of a specific zoning on existingproperty values. The zoning in question is RT-1 which allows conversion to a duplexstructure if the lot size is greater than 6000 square feet. The methodology used is a standardhedonic regression consistent with earlier hedonic studies. The authors attempt to sort outthree potential zoning impacts which are described in Peterson's (1974) working paper onthe influences of zoning regulation on property values. These three impacts are:1) Fiscal impacts arising from the attraction or exclusion of certain activities,2) Impacts from zoning which act as a land use constraint, and3)^Impacts due to externalities as a result of zoning.The results of the study again provide little or no statistical significance of zoning onproperty values.It is interesting that the three studies outlined in the Mark and Goldberg article cometo similar conclusions, yet the methodologies are remarkably different. This has led theauthors to consider why these results may occur. Earlier work typically found similar resultsand concluded that perhaps zoning was not performing the regulatory task it was intended24to. However, Mark and Goldberg determine that perhaps zoning is only one constraint thatmay or may not be binding in a particular market. For example, in New Westminister, themarket for duplex dwellings at the time of the study was slow, which may explain theinsignificant results. This issue must be addressed when interpreting results of any study.An insignificant result may imply that zoning is not a binding constraint. This result hasdifferent implications when addressing public policy issues than the result that zoning orrezoning has no impact on property values.In a subsequent paper, Mark and Goldberg (1986) extend their analysis of theKerrisdale neighborhood in Vancouver using a study period spanning from 1957 to 1980.The methodology used is an hedonic regression using the natural logarithm of actual sellingprice as the dependent variable. Three hypotheses are considered (Mark and Goldberg,1986, p. 260):1) A parcel's zoning classification affects its sale price, and the magnitude anddirection of the effect are consistent over time (a price-effect hypothesis).2) A parcel's sale price is lowered by permitting non-single family uses in theneighborhood (an externality-effect hypothesis).3)^A parcel's sale price is raised by zoning changes which allow higherdensities and different uses, and these effects are consistent through time(a rezoning-effect hypothesis).It is the second hypothesis that is of interest in this current analysis on neighborhood effects.However, all three effects must be considered in the model specification in an attempt toseparate the individual effects. This may be constrained by data availability. The majordeparture in this study is that while previous studies analyzed zoning impacts at one pointin time, the authors consider the long run impacts of zoning.25The data used in this study is obtained from the British Columbia AssessmentAuthority which keeps detailed information on single-family parcels including actual salesprices, transaction histories and an extensive matrix of structural characteristics. Twoneighborhoods are included in this study, Kerrisdale and Fraser. The first neighborhood istypified by a higher affluence than the latter. Both neighborhoods had major rezonings tomulti-family in 1961 with divergent market responses. The Fraser neighborhood had littledemand for multi-family residential, and as a result of the rezoning there was little marketresponse.The results for both neighborhoods provide little support for zoning advocates. TheKerrisdale results only marginally support the price-effect hypothesis, provides some supportfor some of the externalities in the externality-effect hypothesis and offers no support for therezoning effect hypothesis. Results for the Fraser neighborhood provide even less supportfor the three hypotheses. In addition the authors find little or no consistency in the resultsin both neighborhoods, indicating that the externality argument often espoused by zoningadvocates is questionable.The major findings to come from the Mark and Goldberg (1986) study are twofold.First, zoning may not be the panacea to externalities as there may be additional salient forcessuch as consumer preference behavior and other market forces that impact the price ofhousing. Secondly, the cross-sectional methodology used in earlier research comes undersuspicion as the impacts of zoning have a temporal impact. Data that has both a cross-sectional and a temporal component, sometimes referred to as panel data, should be usedwhen studying such phenomenon as rezoning.262.3.2 Additional Research on the Externality ArgumentThere are four additional papers that merit discussion in this section. Although theydo not directly address the zoning issue, they each contribute to the understanding of someof the major issues.Grieson and White (1989) attempt to answer the question of why the current literaturehas found little support for the capitalization of negative externalities in residential propertyvalues. Grieson and White extend the hedonic specification to incorporate a variant on the'agglomeration economy' theme and offer the following model:P = f( W, E, g(N), g(F), g(S) )^ (2.8)where:^g(N) = the probability of conversion in the near future,g(F) = the probability of conversion in the far future,g(S) = the probability of no conversion,E = a measure of neighborhood externalities,W = a vector of other salient characteristics.Their 'agglomeration' argument implies that \"a lot at the edge of an externality producingdistrict has a higher probability of conversion in the future than a lot with no externalityproducing uses in the neighborhood\" (p. 70). Since these probabilities of conversion are notobservable, the authors derive two dichotomous variables, zoning (Z) and externality (E).The control case is stipulated as both being zero. E equalling one and Z equalling zero isthe case of a 'pure externality'. E equalling zero and Z equalling one is the case of a 'pureconversion'. When both equal one, there is a 'combination' effect. The 'pure externality'case implies that conversion is more distant, hence a proxy for g(F). The 'combination' case27implies that conversion is more near, a proxy for g(N). This provides support for the ideathat the impacts of a rezoning are spatial and adjacent properties appear to bear the greatestimpact.The results of their new specification seem to refute earlier reported results regardinglack of capitalization. Grieson and White conclude that previous studies suffer from missingvariable bias. They also add that their proxy measures for the probability of conversion areover simplified. Their methodology has appeal in that it implies that residential propertiesthat are adjacent to a non-residential area may actually appreciate due to the high probabilityof rezoning or conversion in the near future.The effects of high-rise office buildings on residential property values have beenstudied by Thibodeau (1990). His motivation for research on the topic parallels the concernsoutlined in the introduction to this paper, namely that \"public debates are fuelled more byemotions and misinformation than they are by statistical evidence\" (p. 402). An hedonicestimation technique is employed to assess the impacts of an eleven-story office buildingconstructed in North Dallas in 1980. The model breaks down the 'proximity tononconforming uses' variable into two functions, a logarithmic function and a negativeexponential function. A negative sign on the estimated parameter of the logarithmic functionand a positive sign on the estimated parameter of the negative exponential function impliesthat the log component measure the positive benefit and the exponential component measuresthe negative benefit of the proximity to the nonconforming use. Empirically this formulationrequires use of nonlinear least squares and the results of this estimation are as follows:1) Prior to construction, there were no externalities associated with the vacantlot.282)^After construction there were significant externalities, negative forproperties less than 1000 feet from the development and positive forproperties 1000 to 2500 feet from the development.The author concludes that homeowners' concerns regarding reduced property values areindeed justified, but on an aggregate level there are net benefits to home values. If therewere some method of redistributing these gains (ie: through taxation), a more equitablesolution would occur.Michaels and Smith (1990) use hedonic modelling to study the impacts of hazardouswaste sites on housing prices. They use a semilogarithmic form and incorporate threevariables to measure the externality. These are a distance measure and two interactionvariables that attempt to capture short-term response to an announcement such as a rezoningproposal. Their results of concern here are that in the hedonic model 1) marketsegmentation is important and can greatly influence the interpretation of coefficients, and2) introduction of interaction variables may capture important temporal patterns inexternalities.In another paper related to analysing the impacts of toxic waste sites, Palmquist(1992) \"demonstrates that hedonic techniques are particularly well adapted to studyinglocalized externalities since the key problems with the hedonic methodology and benefitestimation do not arise\" (p. 59). The author highlights the fact that the hedonic methodologygenerally requires an estimated hedonic price schedule and an estimated demand or bidfunction. This second step is much more complex both at a theoretical and at aneconometric level. The author argues that the nature of localized externalities allows analysisto stop at the first step. This is because while some residents are affected by the29neighborhood externality, there are many more who are not, which implies that theequilibrium hedonic price schedule remains unchanged. The implication for the currentstudy is that a relatively homogeneous study area be considered when analyzing the impactsof a localized externality.These four papers each contribute to the understanding of estimating the impacts ofexternalities on property values. They can be seen as a sample of the extensive literatureon the topic of using the hedonic methodology to study neighborhood externalities.2.3.3 Higher Density Housing and Property ValuesThis final section takes a specific externality, namely low-income housing, and itseffects on property values. Since low-income housing often implies increased density, it isassumed that the results will act as a proxy for the increased density issue. Of the six papersdiscussed here, only one found negative effects on property values. This enhances thegeneralization that there is statistically little or no evidence supporting the hypothesis thatincreased density lowers adjacent property values. However, other issues aside from thedensity issue may be producing these insignificant results. An increase in crime that is oftenassociated with low neighborhood income or bigotry directed at the poor are difficult toincorporate into a model of property values but may play a significant role. Since previoussections have went into great detail, only a brief summary of the results will be reportedhere.Guy, Hysom and Ruth (1985) study subsidized housing in Fairfax County, Virginia30between 1972 and 1980. They analyze four 'clusters' of townhomes, two near subsidizedhousing, one separated from the subsidized housing by an elementary school and the fourthcluster approximately one half mile from the housing project. An hedonic regression isutilized incorporating a 'distance from subsidized housing' variable from the time the housingprojects were constructed. They find the coefficient on the distance variable to bestatistically significant and positive meaning that the further away from the project, thehigher the value of the house, holding all else constant.Nourse (1963) is evidence of the early work on the subject matter. The author looksfor positive spillover effects of public housing on adjacent housing sites. The study periodis from 1937 to 1959 is St. Louis, Missouri. A repeat sales methodology of constructing aprice index is incorporated into three affected areas and study areas. The price indexes forthe control areas are then compared against those constructed in the affected areas. Theresults indicate that no significant differences are apparent in two of the areas while the thirdarea shows a slight increase in property values.Mitchell-Lama, or The Limited-Profit Housing Companies program provides a liaisonbetween private and public enterprises in the construction of rental housing in New York.In an effort to document the effects of their projects, a study was performed in New YorkCity during 1964-1969 (DeSalvo, 1974). Fifty neighborhoods are studied, eachneighborhood comprising approximately two city blocks. Each borough acts as a controlarea for a number of sites. Comparisons using assessed values are done between theneighborhoods and the control sites. The results indicate that assessed values around therental housing projects increased, on average, five percent higher than in the control areas.31This implies that these projects are actually beneficial to the neighborhoods providingassessed values reflect market values.Rabiega, Lin and Robinson (1984) look at public housing projects in Portland,Oregon between 1963 and 1978. This study focuses on low income whites. This removesthe need to consider ethnic mix, simplifying the analysis. Single family house sales aroundsix public housing projects are used in the analysis. The sales prices are adjusted forinflation and appreciation. Two statistical methods are used on sales before and after theconstruction of a project. These are regression analysis and paired t-tests. The resultsindicate that residential property gained in value after public housing was introduced. Theauthors add that their results may indeed be conservative due to their methods of adjustment.Bab, Pol and Guy (1984) consider the impact of Federally-assisted housing on single-family housing sales in Memphis, Tennessee between 1968 and 1979. Natural neighborhoodboundaries are established around 22 study areas. Control areas as close to the study sitesas possible are identified in order to maintain homogeneity. Regression analysis isperformed using as the dependent variable the ratio of sales prices in the study areas toprices in the city. This ratio effectively adjusts for inflation and provides a check for priceappreciation. The results imply that the introduction of public housing did not cause homesales prices near the housing projects to decline.Farber (1986) studies the impact of group homes for the mentally handicapped in theShreveport-Bossier City, Louisiana metropolitan area between 1979 and 1983. This studyis a departure from the others in this section in that the group homes are not structurallydifferent than the neighboring single-family homes. The local residents did oppose the32introduction of a group home in their neighborhood citing decreased property values as aninherent result. Regression analysis finds no price effects on homes in higher income areasafter the introduction of a group home and also finds positive price effects on homes in lowerincome areas. However, the introduction of a group home in a higher income area increasesthe time-on-market for homes closer to the group home.There is an abundance of literature pertaining to specific urban markets. Sincehousing markets are localized in nature, generalization of any of these studies to any othersub-market would be imprudent.The intent of this literature review has been to determine whether a rezoning wouldaffect adjacent property values, and if so, how can one go about testing for this impact.Both the theoretical and empirical literature indicate zoning affects many aspects of thetypical urban model. Distance from a zoning change boundary has been incorporated intoStull's theoretical model, indicating that econometric analysis of zoning ought to considersuch a measure in the estimating model. This has been the case in many of the empiricalstudies examined in this literature review. Hedonic modelling appears to be a standard toolfor examining housing prices and their determinants. The inclusion of a zoning distancemeasure into a hedonic framework will form the basis of the analysis to follow in the nextchapter.33Chapter 3 The StudyThe hypothesis to be tested in this study is as follows13: \"A parcel's sale price islowered by permitting non-single-family uses in the neighborhood or, in Mark andGoldberg's term, an externality-effect hypothesis\". Refutation of this hypothesis does notnecessarily imply that rezoning has no effect upon adjacent property values. It could alsoimply that the impacts of a rezoning are combined with other non-quantifiable impacts whichare difficult to separate and analyze independently. One divergence from much of the earlierwork is that the study area, the Arbutus Industrial Area, is being rezoned from industrial toresidential, a rezoning that theoretically ought to increase property values in the surroundingneighborhood.In order to undertake such a study , a number of steps are to be considered. Athorough understanding of the neighborhood is essential in order to identify any historicaland/or current concerns that may impact property values. Secondly, a detailed outline of themethodology sheds light on the various factors that ought to be incorporated into a model.Thirdly, the appropriate data must be aggregated and suitable proxies determined to fit themodel. Finally, the model is to be implemented and the results analyzed.13This hypothesis is the second of three that Mark and Goldberg (1986) tested in their studyof the impacts of zoning over time in Vancouver.343.1 Kitsilano: A Brief HistoryThe neighborhood of Kitsilano was born in 1905 with the extension of a streetcar lineover the Kitsilano Trestle Bridge and extending to Balsam Street (refer to Figure 3.1). Fouryears later, another streetcar line was built that extended service along 4 th Avenue to AlmaStreet. The area developed along these service lines which in essence produced twodeveloped areas that later merged into one unified area.During this period of development of the neighborhood, the Kitsilano Ratepayers'Association formed (1906) and it has the honor of being the first association of its kind inCanada. It had as its mandate the responsibility \"to look after the interests of propertyowners through orderly development in Kitsilano, and through keeping in contact withVancouver-wide projects that might affect the area\" (The Vancouver Urban Research Group,1972, p. 69).In 1929, Vancouver adopted a zoning by-law which had as its major explicit goal \"thebelief that the single-family home owner was the backbone of society\" (Gutstein, 1975, p.104). This by-law created three types of residential areas; single-family, two-family andmultiple-family. Since much of the neighborhood was built up before this zoning by-law waspassed, Kitsilano was primarily composed of single-family dwellings.Little happened from this time until after World War II at which time there wasconsiderable construction of three-storey walkup apartments in the eastern part of Kitsilano.It was at this point in time that illegal secondary suites began to appear in single-familyresidences.35.4.493\u00E2\u0080\u00A2`,.Prospect PtFigure 3.1\u00C2\u00B0110,&wash ReaThad BeachFemme PtPool,^Secord BeachPool\u00E2\u0080\u0094AcENGLISHEnghshBayBeachBAYLocamot..4):1ef-\"INT4TTF WI at, 07.7MENA% AVEewePerk; ramolledge^\"'\"b\u00E2\u0080\u00A2^KITS:' BROADWAYFAIRVIEW:7,SCAAirry'. agVancoPrefadaw!toSpital \u00E2\u0080\u00A2N DS:KIN ED WAJESSYT 2TS:07704^\u00E2\u0080\u00A24c.,,al^\u00E2\u0080\u00A2**siAVE:e7CC\"*ST43,v1P1^'NMalkinPk \u00E2\u0080\u0094 &CO Aiwa Sscom.A.od a z- 47 Awl \"-14., .^3^ATIMon NI ornary \u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2.:Park 41 ;,..17, ...A44 *47PeAyiSS 004\"AVE Ookne;SPO0002, a r*C4 -436In an attempt to eliminate these illegal secondary suites, the City of Vancouver in1951 aimed its attention at the Kitsilano neighborhood. The Kitsilano ImprovementAssociation (KIA) formed as an interest group maintaining that these secondary suites werenecessary for housing the population of the neighborhood. They fought for rezoning to two-family residences and in 1959 a new zoning by-law was approved that would in essencemake these illegal suites legal in certain areas of the neighborhood. Amendments wereadded in 1961 which was the construction of several highrises on the lower slopes ofKitsilano between 1962 and 1968.Developers were busy buying up the transitional properties which were zoned two-family in anticipation that as growth continued, these properties would be rezoned formultiple-family residential. They rented out these speculative properties to a less stablepopulation. It is interesting to note that at this time in history, the general feeling was thatKitsilano would eventually be all multiple-family residential which would then put pressureon areas such as Point Grey and Dunbar. This view is consistent with the 'growthperspective' which was common at the time.By 1971, the population of Kitsilano had reached 34,000 which was only a sixpercent increase from the population in 1951. In this same time period, the population ofVancouver as a whole increased 25 percent making Kitsilano one of the slowest growth areasin the city. This slow growth can be attributed to the change in the demographic structureof the neighborhood. Once a common area for families with children, Kitsilano, perhaps dueto the pressures of central city development, became a haven for the 20-to-34-year-oldgroup.37The Kitsilano Area Resources Association (KARA) had recently formed by this timeand were busy protecting the neighborhood from transportation improvements that wouldhave seen the area severed. This interest group took on a proactive stance as they becameinvolved in a diverse set of issues including alternative housing schemes for theneighborhood, fighting the 'almighty' highrise, beautification of Broadway Avenue, theinfamous third crossing, and improved public transportation for the area.Another interest group stood up to the redevelopment of the neighborhood in 1974when a developer proposed razing seven houses on one block to make was for a three-storeycondominium apartment. The West Broadway Citizens Committee (WBCC) stood betweena house and the bulldozer as the residents of the house pondered their legal entitlements.The developer won this fight in the end but not without bringing some interesting mediacoverage. The 1970's was a decade that was witness to high growth in many cities in NorthAmerica, which led to the boom of neighborhood groups that were created in response tothis unprecedented growth.3.2 Kitsilano Today: The Arbutus Industrial AreaIt has been determined that by 1992, the population of Kitsilano had dropped byseveral thousand from its 1952 level'''. The continued high growth rates that the GreaterVancouver region was experiencing was prompting civic officials to determine alternative\"The Vancouver Sun, Monday February 10, 1992, page A13. \"Arbutus Lands: ObjectorsWould Price People out Completely\". The demographic changes that occured in the neighborhoodover the past two decades resulted in an overall decline in population. These demographicchanges include the transition of the neighborhood from family orientation to an enclave for thesingle working class.38solutions to alleviate the already tight housing market. The Planning Department was askedto identify specific area in Vancouver that would be suitable to redevelopment into higherdensity residential. In the fall of 1989, Vancouver City Council identified the opportunityof transforming the Arbutus Industrial Area (AIA) into much needed housing. Up to 1300new housing units would be realized in an area bounded by Vine Street, Broadway Avenue,Maple Street, and West 12th Avenue.By December of that year, the Kitsilano citizen's Planning Committee (KCPC)formed with a mandate to reach a consensus with city and developers. The main issue wasdensity of the Arbutus site. Four months later, the property owners of the retired MolsonBreweries site presented a scheme for the area which included high density development,Floor Space Ratio (FSR) of 3.45 and building heights reaching 13 stories.KARA (now an abbreviation for Kitsilano Arbutus Resident's Association) voicedtheir opinion in May 1991 by stating that the neighborhood could not support densities higherthan 1.5 FSR and building heights of no higher than four stories. The City PlanningDepartment indicated trigger density ratios of 1.8 to initiate redevelopment and 2.3 toprovide social housing, park space and other amenities.In November 1991, yet another interest group emerged. It was called the ArbutusVine Improvement Group (AVIG) and represented the interest of the developers of theindustrial area. Their high quality pamphlet, which was distributed to residents in theneighborhood, contained impressive pictures of what high density redevelopment would looklike. KARA immediately took a stance and accused the AVIG of trying to subvert theplanning process.39By January of 1992, the City Planning Department released its proposals andrecommendations for the AIA in a public meeting in the Kitsilano High School There wassuch a tremendous turnout of concerned citizens that three additional meetings werescheduled in order to hear all the concerns. One of the arguments used by KARA was thatthe existing infrastructure would not accommodate such a grand redevelopment scheme. Tomake this point clear to the Kitsilano residents and to the residents of Vancouver as a whole,members of KARA drove around the AIA protesting the increased congestion and thuscreating an environment which they thought would transpire were the redevelopment toproceed.One month later yet another interest group emerged, one representing the existingland owners and labelling themselves the Arbutus Industrial Area Landowners Group(AIALOG). It was this group that accused the KCPC of playing the YIMBY-BUTNOT (YesIn My Back Yard, BUT NOt like This) game. Instead of implying that other areas inVancouver submit to redevelopment, the KCPC welcomed the AIA proposal, 'on theirterms'. These terms would make the redevelopment scheme infeasible for any developer andthus the NIMBY results would be attained without succumbing to the negative connotationsof protecting self interests. AIALOG tried to argue that the AIA was a 'distinct'neighborhood from Kitsilano, perhaps trying to disqualify the arguments of KARA andKCPC.In March 1992, City Council rejected the moderate density proposal submitted by thePlanning Department stating that the neighborhood would be 'adversely affected by theredevelopment plan'. A motion was also passed which effectively reduced the community40amenity contributions and dropping the amount of social housing necessary on the site in anattempt to stimulate the revised redevelopment scheme. It seems ironic that reducing theamount of social housing contradicts their original mandate of finding affordable housingalternatives for the residents of Vancouver.Nearly four months later, City Council approved a redevelopment scheme whichwould realize building heights with maximum heights of eight stories. Both residentassociations felt that the City had not listened to their concerns and that the Non-PartisanAssociation (NPA) dominated council was siding with the corporate interests. As of latesummer, 1993, the rezoning issue has yet to be finalized.3.3 MethodologyThe method chosen to test the 'externality-effect' hypothesis is the hedonic approachdiscussed extensively in the literature review. In simple terms, this approach considers theattributes that significantly affect the selling price of a parcel of land. Such attributes includethe size of the lot, and the age and size of the improvements (structure). The variable ofinterest, distance from the AIA rezoning, is incorporated into the model and tested todetermine whether it significantly impacts the selling price.There have been many functional forms of the hedonic price function deemedappropriate for use in studying housing price impacts (Hamilton, 1992) but the one chosenfor this study is based in the Cobb-Douglas framework. This framework is used as it hasbeen found that the various housing characteristics used in the hedonic studies have a41diminishing effect upon housing prices. For example, the marginal impact of a thirdbedroom is intuitively higher than the marginal impact of a sixth bedroom in thedetermination of housing prices.Equation 3.1 summarizes the basic model.RPricei^Ca11131X21132X13X413481^ (3.1)where:^RPricei =^real price of household,Xli^=^vector of continuous structural attributes forhousehold; (for example, number of bedrooms),X2i^vector of dichotomous structural attributes forhousehold; (for example, garage: O=no, 1 =yes),X31^vector of neighborhood attributes for householdbX4i^=^vector of zoning effects for household,a, 13,c1^ constant, coefficients and error term respectively.As this study is confined to one homogeneous neighborhood' the vector of neighborhoodattributes can be removed from the basic model. By taking the log of both sides of thismodel, a linear form is constructed suitable for estimation using ordinary least squares'.As prices are typically collected in nominal terms, a series of quarterly dichotomousvariables are added to the model to identify the local price trend. The regression model isthus constructed in the following manner.= a + 1311nX1i + 132X21 + 1341nX41 + tisai + ci^(3.2)For model simplification, only one neighborhood is used in this study. This takes care ofthe problem of estimating coefficients that are considered unique to each particular neighborhood.The trade off is that the sample size is rather limited.'Ordinary Least Squares (OLS) is the most common method of estimating a linear equation.For details of the estimation procedure, any introductory statistics text is suitable.42where:^1nPi =^log of nominal price of householdi,=^vector of log of continuous structural attributes forhouseholdi (for example, ln(number of bedrooms)),X2i^vector of dichotomous structural attributes forhousehold, (for example, garage: 0 =no, 1 =yes),lnX41 =^vector of log of zoning effects for householdi,=^vector of dichotomous quarterly variables used toidentify a local price trend,af Pp ei^constant, coefficients, and error term respectively.The estimated coefficients in this functional form are read as elasticities. In other words,if the estimate of f3 for lot size is 0.006, a one percentage increase in lot size, holding allelse constant, would imply a 0.6 percent increase in the price of the house. Caution is inorder in interpreting the dichotomous variables as a mathematical transformation is requiredbefore reading their coefficients as elasticities'.3.4 The DataThe vector of structural attributes required for a complete specification of the hedonicprice equation would be costly in terms of data collection. However, previous hedonicmodelling studies have determined that a select number of variables prove sufficient forestimation. Table 3.1 lists the variables that are used in this study.The dependent variable used in the hedonic model is price which is estimated in bothnominal and real dollars. This is the transaction price that is reported to the Land Titles\"The coefficient of a dichotomous variable in a log model is a percent change in price but thepercent is (e\u00C2\u00B0 - 1) rather than f3. For example, if the coefficient is estimated to be 0.03, thepercentage increase is (e\"3 - 1)x100, or 3.04 percent.43Office. The year and month variables coincide with the closing date of the sale. Theremaining variables require no additional explanation.The British Columbia Assessment Authority (BCAA) maintains a database of salesactivity in the province. This database, in conjunction with information collected by theMultiple Listing Service (MLS) allow the construction of a large and comprehensive crossindex maintained by Stan Hamilton at the University of British Columbia. The data for thisstudy are taken from these combined files.TABLE 3.1Variable^lagPrice (in nominal dollars)^ContinuousYear soldMonth soldAge of house (years)Lot width (feet)Lot depth (feet)Main floor area (square feet)Number of bedroomsNumber of other roomsNumber of basement roomsNumber of fireplacesNumber of full bathsNumber of partial bathsGarage^ DichotomousPresence of sewerCurb & gutterPresence of sidewalkCorner lotPrime viewIn addition to the variables outlined above, a set of quarterly dichotomous variablesare used to capture the temporal nature of the data. Alternatively, an hedonic price index44for single detached housing in the city of Vancouver is used to deflate the nominal housingprices, thus removing the time trend. Over the short study period, it is assumed that theother variables are time insensitive.The zoning variables added to the list include DIST, a distance measure (in metres)between each individual parcel the AIA and EXP, an indicator variable separating the studyarea data from the control area data. The distance variable is the euclidean distance betweeneach property and the proposed rezoned area of the Arbutus Industrial Area. This directmeasure of distance is preferred over another more indirect measure of following the existingroadway network as the diseconomies associated with this particular rezoning includeexcessive noise levels and blocking of the views of the north shore mountains, both beinguncorrelated with the defined street layout.3.5 The Study AreaThere are two reasons for choosing this particular rezoning for analysis. The first isdue to the close proximity to single family residential zoning. The second is that themagnitude of the rezoning and the public reponse make it an attractive study area. The northboundary of the study area is West 12th Avenue which also constitutes the south boundaryof the proposed Arbutus rezoning. The study area extends south to the lane separating West15th Avenue and West 16th Avenue and lies between Arbutus Street on the east and LarchStreet on the West. The study area is comprised of a total of 266 residential homes, 12commercial buildings and three multi-family complexes. This area can be divided equally45into two segments, seven blocks immediately south of the Arbutus site and seven blocksdirectly south of Connaught Park (refer to Figure 3.2).Of the 83 actual sales that occurred in the study area (refer to Table 3.3) over the1987 to 1992 time period, 69 sales occurred in the area directly south of Connaught Parkwhile only 14 sales occurred in the area directly south of the Arbutus site. Two explanationscan be offered for this unusual distribution of sales. The first relates to speculative activitywhile the second relates to the social mix of the residents. With a rezoning imminent, salesactivity may slow down pending the outcome of the process, a result of speculative activity.This result is most common in transition zones thatTABLE 3.2 SUMMARY STATISTICSCombined Study Area and Control Area (n=388)Variable Mean Std Dev Minimum Maximum LabelNOM. PRICE 303164.38 97734.52 156000 695000 NOMINAL PRICEYEAR 89.43 1.71 87 92 SALE YEARMONTH 6.69 3.21 1 12 SALE MONTHAGE 51.18 17.01 0 83 AGE (years)WIDE 38.39 7.43 24.5 61.0 LOT WIDTHDEEP 122.17 4.43 66.0 125.0 LOT DEPTHGARAGE .83 .38 0 1 GARAGEBEDROM 3.70 1.42 1 8 NUMBERSEWER .92 .27 0 1 SEWERCURB .94 .24 0 1 CURBSIDEWALK .94 .24 0 1 SIDEWALKCORNER .17 .37 0 1 CORNERMAIN 1161.44 314.78 476 2267 MB FLR AREAFIRE 1.08 .45 0 3 FPL #BASEROM 1.01 1.18 0 4 BASEMENTOROOMS 4.83 1.84 2 11 TOTAL ROOMSFBATHS 1.45 .72 1 5 FULL BATHSPBATHS .43 .58 0 3 PART BATHSEXP .21 .41 .0 1.0 EXPERIMENTDIST 950.67 381.88 45.0 1400.0 DISTANCE46\u00E2\u0080\u00A2 WW, w8\u00E2\u0080\u00A2 \u00E2\u0080\u00A2\u00E2\u0080\u00A2 \u00E2\u0080\u00A2AVE. 6 8 6cold.69010 TH.IOTSILANO,R. S.L ORTENN)E.AVE01111 411111I vasONP MID SO - AIM ea^vim 11111013 D-114=0Pit 14 TH.15 TH,6wo4^2400AVE.^654 I670.1.41111\u00E2\u0096\u00A0 41111. 411111 68303AVE.ta41..14^\u00E2\u0080\u00A2 \u00E2\u0080\u00A2 \u00E2\u0080\u00A2 \u00E2\u0080\u00A2typically circle the central business district. A characteristic of these transition zones isan obvious lack of structural maintenance. However, visual inspection indicates that thearea with low sales activity is more stable with larger, well maintained homes and yardswhich may contradict the speculative theory. It also seems unlikely that there could beenough social diversity in such a small neighborhood to create the unbalanced distributionof sales. Although both theories have their merits, realtors familiar with theneighborhood could only offer an explanation to reinforce the stability hypothesis\".3.6 The Control AreaThe control area used in this study is situated in the same neighborhood thushopefully mitigating extraneous neighborhood effects into the analysis. Ideally, the areamust be far enough away from the Arbutus Industrial Area so as not to capture theexternality effect of the rezoning. The area chosen is situated west of Macdonald Street,east of Blenheim Street, and has the same north and south boundaries as the study area(refer to Figure 3.3). Of the 383 homes that occupy the control area, there are a total of305 sales, a number almost four times as great as the sales activity in the study area.Again, the reasons discussed above regarding the distributional disparities in the studyarea can be again referred to.'This phenomenon was brought to the attention of two realters that are familiar with theneighborhood. Both found the results unusual but could not provide an alternative explanation tothe ones offered above in the text. One realtor stated that he made cold calls in the area andfound that there are many older home owners that are just not interested in selling their properties.48^ 1 L..1^t\u00E2\u0080\u0094JFigure 3.3r--cati\u00E2\u0096\u00A0\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0096\u00A0el VDFL 1VAs&DI-0IIIM.1!\u00E2\u0096\u00A0\u00E2\u0096\u00A0\u00E2\u0096\u00A01.1\u00E2\u0096\u00A01.\u00E2\u0096\u00A0\u00E2\u0096\u00A0\u00E2\u0080\u00A2.\u00E2\u0096\u00A0\u00E2\u0096\u00A0\u00E2\u0096\u00A0\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0096\u00A011,.1\u00E2\u0080\u00A2=1M.=\u00E2\u0096\u00A0\u00E2\u0096\u00A0\u00E2\u0096\u00A0\u00E2\u0096\u00A0\u00E2\u0096\u00A0\u00E2\u0096\u00A0OCO 3H\u00E2\u0080\u0098 N319.11\u00E2\u0096\u00A0\u00E2\u0080\u00A2\u00E2\u0096\u00A0\u00E2\u0080\u00A2\u00E2\u0096\u00A0\u00E2\u0080\u00A2\u00E2\u0080\u00A2 \u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0096\u00A0\u00E2\u0080\u00A2\u00E2\u0096\u00A0\u00E2\u0080\u00A2\u00E2\u0096\u00A0\u00E2\u0080\u00A2\u00E2\u0096\u00A0\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0096\u00A0\u00E2\u0096\u00A0\u00E2\u0080\u00A2 \u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0096\u00A0\u00E2\u0096\u00A0\u00E2\u0096\u00A0\u00E2\u0080\u00A2\u00E2\u0080\u00A2J. S49\u00E2\u0080\u00A2\u00E2\u0096\u00A0\u00E2\u0096\u00A0\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0096\u00A0\u00E2\u0096\u00A0\u00E2\u0096\u00A0\u00E2\u0080\u00A2InS...\u00E2\u0096\u00A0\u00E2\u0096\u00A0\u00E2\u0096\u00A0\u00E2\u0080\u00A2110\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0096\u00A0\u00E2\u0096\u00A0\u00E2\u0080\u00A2\u00E2\u0080\u00A2\u00E2\u0080\u00A2.1.111\u00E2\u0096\u00A0\u00E2\u0096\u00A0\u00E2\u0096\u00A0\u00E2\u0080\u00A2TABLE 3.3 SUIVIIVIARY STATISTICSSTUDY AREA (n=83)Variable Mean Std Dev Minimum Maximum LabelNOM. PRICE 354932.53 116612.17 160000 695000 NOMINAL PRICEYEAR 89.66 1.65 87 92 SALE YEARMONTH 6.89 3.41 1 12 SALE MONTHAGE 48.61 18.73 1 83 AGE (years)WIDE 47.58 7.42 25.0 50.2 LOT WIDTHDEEP 125.00 .00 125.0 125.0 LOT DEPTHGARAGE .83 .38 0 1 GARAGEBEDROM 3.76 1.45 1 7 NUMBERSEWER .82 .39 0 1 SEWERCURB .82 .39 0 1 CURBSIDEWALK .82 .39 0 1 SIDEWALKCORNER .13 .34 0 1 CORNERMAIN 1343.49 418.89 476 2267 MB FLR AREAFIRE 1.19 .61 0 3 FPL #BASEROM 1.13 1.27 0 4 BASEMENTOROOMS 5.64 2.02 3 11 TOTAL ROOMSFBATHS 1.65 1.10 1 5 FULL BATHSPBATHS .52 .61 0 2 PART BATHSEXP 1.00 .00 1.0 1.0 EXPERIMENTDIST 280.24 97.79 45.0 550.0 DIST. FROM AIACONTROL AREA (n=305)Variable Mean Std Dev Minimum Maximum LabelNOM. PRICE 289076.65 86955.12 156000 611215 NOMINAL PRICEYEAR 89.48 1.69 87 92 SALE YEARMONTH 6.64 3.15 1 12 SALE MONTHAGE 51.88 16.47 0 82 AGE (years)WIDE 35.89 5.10 24.5 61.0 LOT WIDTHDEEP 121.40 4.71 66.0 122.4 LOT DEPTHGARAGE .83 .38 0 1 GARAGEBEDROM 3.69 1.42 1 8 NUMBERSEWER .95 .22 0 1 SEWERCURB .97 .17 0 1 CURBSIDEWALK .97 .16 0 1 SIDEWALKCORNER .18 .38 0 1 CORNERMAIN 1111.90 259.42 624 1780 MB FLR AREAFIRE 1.05 .39 0 3 FPL #BASEROM .98 1.16 0 4 BASEMENTOROOMS 4.61 1.73 2 10 TOTAL ROOMSFBATHS 1.39 .56 1 4 FULL BATHSPBATHS .40 .57 0 3 PART BATHSEXP .00 .00 .0 .0 EXPERIMENTDIST 1133.11 164.19 900.0 1400.0 DIST FROM MA503.7 Discussion of the Study Area and Control AreaThe characteristics of the combined study area and control area are outlined in Table3.2. Table 3.3 provides summary statistics of each separate area for comparison purposes.The average sales price of a single family home during the six year study period is $355,000in the study area as compared to $290,000 in the control area. This difference in values canbe explained be observing the summary statistics. The age of structures in the study areaare, on average, 49 years compared to 52 years in thecontrol area. Although the study area appears to have more older homes, there have beenmany renovations in the area which would explain the younger average age. The homes inthe study area are, on average, on larger lots, have more bedrooms and other rooms, andencompass larger floor plates. The study area is characterized as having less infrastructurewith only 82 percent of the sales having such amenities as curb and gutter, sidewalks andsewer. The comparable number in the control area is 97 percent.The distance from the proposed rezoning averages 280 metres for the study area withthe closest property being 45 metres and the farthest property being 550 metres. The controlarea has a minimum distance of 900 metres and extends to 1400 metres away from theproposed rezoning. It is anticipated that this distance and the fact that the control area isseparated from the Arbutus Industrial Area by a major arterial roadway (Macdonald Street)will minimize the impact of the rezoning on the control area.Both the study area and the control area have similar access to the central businessdistrict (CBD) and to the Broadway corridor. Both are situated near a major park with the51study area adjacent to Connaught Park and the control area adjacent to Carnarvon Park. Theschools in the area are evenly distributed between the areas. This implies that incorporatingsuch variables as access to the CBD and proximity to parks and schools can be omitted fromthe analysis, simplifying the model.3.8 Regression ResultsOrdinary least squares regression was run using the model discussed in Equation(3.2). The results are reported in Table 3.4. The highly significant coefficients on thequarterly indicator variables indicate that there is a substantial temporal trend in the housingprices. This result is not surprising. Aside from the quarterly variables, there are fivevariables that are significant at the ten percent level of significance. These five variables alldisplay the expected signs and the coefficients are comparable to other studies of housingdata in the lower mainland (Hamilton and Carruthers, 1993). The zoning variable reportedis the log of the distance measure (LNDST). The coefficient is not significant and thenegative coefficient contradicts the assumed externality effect. A second regression was runsubstituting the dichotomous zoning variable for the continuous measure. The results do notvary substantially.Residual analysis indicated the presence of two outliers in the data. An investigationindicated these observations were not errors in the data and were thusnot removed from the data set. Aside from the outliers, the residuals appear homoscedasticand thus no corrections have been made to the model.52One of the results that was to be determined from the analysis was whether there hasbeen a change in the zoning impact due to the announcement of the rezoning in the fall of1989. It is assumed that in an efficient land market, the announcement of a possiblerezoning would immediately be capitalized into the price of the surrounding land.In order to investigate this announcement effect a second model was constructed withtwo major departures from the first model. This second model is formulated as follows,lnDefPricei = a + a'POST +^+^+ 132x21132'POSTX21 + 1341nX41 + 134'POSTInX41 + c,^ (3.3)where:^lnDefPricei^log of deflated price of householdi,vector of log of continuous structural attributesfor household; (for example, ln(number ofbedrooms)),X2i^ vector of dichotomous structural attributes forhousehold; (for example, garage: O=no, 1 =yes),lnX4i^vector of log of zoning effects for householdpPOST Indicator variable (=1 after announcement),fip^=^constant, coefficients, and error termrespectively.The quarterly indicator variables had to be removed from the model and this requireddeflating the nominal prices with some measure of inflation. One such deflator, theconsumer price index for housing in Vancouver as reported in the Cansite database,proved to be a poor indicator of the temporal nature of the data. An additional two priceindexes constructed from single family sales in Vancouver were also used to deflate the'9Statistics Canada maintains an electronic database titled Cansim.53nominal price. It was found that the hedonic price index based on sales in the City ofVancouver proved to be the best indicator of temporal price changes.A differential intercept and a series of differential slope coefficients (Gujarati, 1988,p.446) were added to the model incorporating a new indicator variable (POST) which equalszero before the announcement and one after the announcement. The variables LNBASERMand LNRMS were deleted from the second regression as they were suspected of producingmulticollinearity.The second regression results are outlined in Table 3.5. By deflating the nominalprices the time trend is ideally eliminated. An initial regression incorporating the quarterlyvariables indicated that the hedonic index performed well in removing the temporal natureof the data.Four variables were found to be significant in this regression with an adjusted R 2 of66 percent. These variables include the log of the age of the structure, the log of the mainfloor area, the log of the lot size and the log of the number of fireplaces. In addition, thedifferential slope coefficients for age and garage are significant indicating that there appearsto be some structural change occurring for these variables after thefall of 1989. It appears that the negative impact associated with age has increased since1989 while the presence of a garage now has a significant impact on price.54TABLE 3.4 REGRESSION RESULTSMultiple R^.94611^Analysis of VarianceR Square^.89512 DF^Sum of Squares^Mean SquareAdj. R Square .88502^Regression^34^34.85160^1.02505Standard Error .10755 Residual 353^4.08330 .01157F =^88.61497^Signif F = .0000Dependent Variable.. LNP^Variables in the Equation^Variable^B^SE B^T^Sig TLAGE^-.11741^.00873^-13.443^.0000LNLT .37813 .04277 8.840 .0000LNMAIN^.11133^.02795^3.983^.0001LNBD .04970 .01773 2.803 .0053LNBASRIVI^.01765^.01059^1.667^.0965LNDST^-.0060 .01082 -.559 .5766GARAGE^.01302^.01546^.842^.4001CORNER^-.00170 .01546 -.110 .9124LNFIRE .02247^.0302^.744^.4574LNRMS^.00100 .02267 .044 .9648LNTBATH^.0022^.01889^.117^.9068Q2^.01901 .05054 .376 .7070Q3 .09267^.05051^1.835^.0674Q4^.08965 .05605 1.599 .1106Q5 .14763^.05376^2.746^.0063Q6^.27235 .05007 5.440 .0000Q7 .28595^.05340^5.355^.0000Q8^.32759 .05359 6.113 .0000Q9 .47079^.04991^9.432^.0000Q10^.58957 .04966 11.873 .0000Q11 .54468^.05011^10.871^.0000Q12^.60818 .04999 12.166 .0000Q13 .70054^.05062^13.840^.0000Q14^.63406 .06594 9.616 .0000Q15 .48974^.05537^8.844^.0000Q16^.55341 .05300 10.441 .0000Q17 .58486^.06064^9.645^.0000Q18^.63615 .04896 12.994 .0000Q19 .74590^.05429^13.740^.0000Q20^.76130 .05233 14.547 .0000Q21 .74592^.05544^13.455^.0000Q22^.82449 .05219 15.797 .0000Q23 .84066^.05112^16.444^.0000Q24^.86583 .05020 17.248 .0000(Constant)^8.47670^.36268^23.372^.000055As has been found in all the regressions run, the DIST variable is statisticallyinsignificant at the ten percent level, both before 1989 and after 1989. It appears from thisresult that the distance separating individual properties from the AIA has not had any impacton the value of those properties. This result is consistent with the vast majority of studiesmentioned earlier in the literature review. In subsequent regressions substituting thedichotomous EXP variable for the DIST variable, similar results were obtained. Theseresults tend to refute the tested hypothesis and imply that a parcel's sale price is not affectedby permitting non-single-family uses on the previously zoned light industrial land.The results can also be interpreted in a number of other ways. It is quite possiblethat the distance variable is capturing more than one externality. If it is capturing thepositive externality of rezoning from industrial to residential, and the negative externality ofincreased density relative to adjacent densities then the two opposing results may becancelling each other out.A final interpretation offered in this analysis is the fact that the rezoning has not beenfinalized at the time of this writing. Perhaps the announcement effect will be morepronounced when the official rezoning occurs as opposed to when it was first announced.This interpretation is rather weak and is discussed further in the concluding remarks.56Table 3.5 Regression ResultsMultiple R^.82634R Square .68284Adjusted R Square .66826Standard Error^.10911Analysis of VarianceDF Sum of Squares^Mean SquareRegression^17^9.48380^.55787Residual^370^4.40505 .01191F =^46.85810^Signif F = .0000 ^Variables in the Equation ^Variable B SE B T Sig TLAGE -.09360 .01204 -7.773 .0000LNLT .35070 .06473 5.418 .0000LNMAIN .14024 .04497 3.119 .0020LNFIRE .08384 .04562 1.838 .0669LNBD .03627 .02543 1.426 .1546LNTBATH .02473 .02518 .982 .3268GARAGE -.00896 .02140 -.419 .6754LNDST -.01139 .01733 -.657 .5115POST .19507 .70469 .277 .7821PLAGE -.04615 .01689 -2.733 .0066PLNLT .02863 .08498 .337 .7364PLNMAIN -.05153 .05702 - .904 .3667PLNFIRE -.08238 .05965 -1.381 .1681PLNBD .03655 .03549 1.030 .3037PLNTBATH -.04349 .03459 -1.257 .2095PGAR .05475 .03066 1.786 .0750PLNDST .00640 .02201 .290 .7717(Constant) 3.93785 .51286 7.678 .0000Chapter 4 ConclusionThere are many factors that impact property values in any particular jurisdiction. Themore obvious ones include distance from the central business district and the size of theparcel of land. Structural attributes such as floor area and age have also been deemedsignificant contributors to the total value of the property. These and many others have beenquantified in hedonic studies which have attempted to identify the determinants of value. Inaddition to these overt attributes there are numerous covert attributes which also have animpact on property values. Such neighborhood externalities including traffic congestion,noise and various adjacent land uses have been topics of academic study. The results ofthese studies have varied, with some showing significant impacts on property values relatedto proximity to a particular externality and others showing no significant impacts.Local area interest groups have often used the negative impacts of neighborhoodexternalities in their arsenal of defenses to fight against rezoning applications. The speedand intensity in which interest groups form against any particular rezoning applicationappears to be correlated to the sense of neighborhood and the history of interest groupactivity. Chosen for this study is the Arbutus Industrial Area which is situated in theKitsilano neighborhood in Vancouver's west side. There is a strong sense of communitywhich goes as far back as the formation of the neighborhood earlier this century. Thedecision to rezone the industrial lands into medium density residential has met with a greatdeal of opposition from two prominent interest groups. Although they have beenunsuccessful of their optimal choice of single-family residential for the area, they have seen58a reduction in the overall density slated for the area. At the time of this writing, the finaldensity of development was still undecided.Statistical analysis was performed on the neighborhood surrounding the ArbutusIndustrial Area to determine whether the rezoning has had any impact on adjacent propertyvalues. The method chosen, the hedonic approach, has been used widely in the area ofexternality research.Results of statistical analysis indicate that the distance variable is insignificant. Thisresult can be interpreted in many ways. One interpretation would be that the rezoning hashad no impact on adjacent property values. This interpretation has a significant amount ofresearch to back it up.A second interpretation is that the distance variable is measuring more than oneexternality. For example, the Arbutus Industrial Area is being rezoned from light industrialto residential. Intuitively this conversion to a more compatible land use ought to positivelyaffect adjacent property values. However, the increased density relative to the surroundingneighborhood can be considered a negative externality. The combination of these twoexternalities may be producing the insignificant results.A final interpretation is a result of the current status of the rezoning application. Itcan be argued that the market is confident the rezoned density will not exceed the levelconsidered appropriate for the area. However, the market is fully aware that rezoning isimminent and the density is going to be higher than adjacent densities. If the market wereefficient, property values should already reflect these changes and the legal rezoning datewould not be considered a significant factor.59In the introduction, three cases were addressed regarding neighborhood oppositionto increased density in their neighborhoods. One case introduced the Arbutus area and theneighborhood concerns regarding the rezoning. This case has been the focus of this paper.The second case involved a neighborhood in the City of Coquitlam. The OxfordHeights residents fear that the subdivision of one lot into two smaller lots in theirneighborhood would cause their property values to plummet. The results from this paperappear to counter their fears. This specific case involves no land-use change and requiresno changes to the existing zoning regarding minimum lot size. This case appears to be anexample of the residents fears of neighborhood change and their use of 'declining propertyvalues' to legitimize these fears.The third case articulated the concerns of residents in one of Vancouver's east sideneighborhoods regarding the introduction of a multi-family development on property ownedby a local church. Again, the results here appear to indicate that their property values willnot be affected by the proposed rezoning. In addition, the results from the City ofVancouver's 'New Neighbours' document (City of Vancouver, 1986) indicate that the overallneighborhood concerns regarding a rezoning of this nature appear to subside after the higherdensity has been introduced into the neighborhood.Statistical analysis may appear to be a harsh and cold method of determining whetheror not residents fears of change in their neighborhoods are in fact true. However, themethod is quantitative and the results are based on factual data. A more complete study ofthe problem addressed in this paper would involve the introduction of a qualitativecomponent. Interviewing neighbors both before and after a rezoning, similar to the method60used by the City of Vancouver in their study on neighborhood change, would complimentthe quantitative research of this study. This more balanced approach would address theconcerns of all residents as opposed to focusing on overall impacts on a neighborhood.As cities such as Vancouver evolve and the inner industrial areas become out-dated,there will be a desire to rezone them to more appropriate uses. In an effort to reduce urbansprawl, civic officials will rely more heavily on increased density. This may mean thatpeople and neighborhoods are going to have to adapt to their ever changing surroundings.Existing residents have adapted in the past and they will adapt in the future.61ReferencesAvrin, M.W., (1977), Some Economic Effects of Residential Zoning in San Francisco, in\"Residential Location and Urban Housing Markets\" (G.K. Ingram, Ed.), BallingerPublishing, Cambridge.Babb, Carol, Louis Pol and Rebecca Guy, (1984), \"The Impact of Federally-AssistedHousing on Single-Family Housing Sales: 1970-1980\", Mid-South Business Journal,p.13-17.Bailey, Martin, J., (1969), \"Note on the Economics of Residential Zoning and UrbanRenewal\", Land Economics, 36, p.288-292.Bird, Richard M. and Enid N. Slack, (1983), Urban Public Finance in Canada, Toronto:Butterworth & Co. (Canada) Ltd.Baxter, David, (1989), \"Population and Housing in Metropolitan Vancouver, ChangingPatterns of Demographics and Demand\", Laurier Institute, Vancouver.Brigham, (1965), \"The Determinants of Residential Land Values\", Land Economics, 41,p.325-334.Cheng, Fook-Jam, (1992), \"Neighbourhood Externalities\", Commerce 604 Term Paper,Faculty of Commerce and Business Administration, University of British Columbia.City of Vancouver, (1986), New Neighbours: How Vancouver's Single-Family Residents feelabout Higher Density Housing, Vancouver: Vancouver Planning Department.Courant, Paul, N., (1976), \"On the Effects of Fiscal Zoning on Land and Housing Values\",Journal of Urban Economics, 3, p. 88-94.Crecine, J., 0. Davis and J. Jackson, (1967) \"Urban Property Markets: Some EmpiricalResults and their Implications for Municipal Zoning\", The Journal of Law andEconomics, 10, p. 79-99.DeSalvo, J.S., (1974), \"Neighborhood Upgrading Effects of Middle-Income HousingProjects in New York City\", Journal of Urban Economics, 1, p. 269-277.Epple, Dennis, Thomas Romer, and Radu Filimon, (1988), \"Community Development withEndogenous Land Use Controls\", Journal of Public Economics, 3, p. 133-162.62Farber, Stephen, (1986), \"Market Segmentation and the Effects on Group Homes for theHandicapped on Residential Property Values\", Urban Studies, December, p. 519-525.Fischel, William A., (1978), \"A Property Rights Approach to Municipal Zoning\", LandEconomics, 54, p. 64-81.Gapfert, G., and R. Knight, (1982), \"Cities in the 21st Century\", edited by Gary Gappert andRichard Knight, Sage Publications, Beverly Hills.Grieson, R.E., and J. White, (1989), \"The Existence and Capitalization of NeighborhoodExternalities: A Reassessment\", Journal of Urban Economics, 25, p. 68-76.Goldberg, M. and P. Horwood, (1980) \"Zoning: Its Costs and Relevance for the 1980's\", TheFraser Institute, Vancouver.Grether, D.M. and P. Mieszkowski, (1974), \"Determinants of Real Estate Values\", Journal ofUrban Economics, 1, p. 127- 146.Grether, D.M. and P. Mieszkowsld, (1980), \"The Effects of Nonresidential Land Uses on thePrices of Adjacent Housing: Some Estimates of Proximity Effects\", Journal of UrbanEconomics, 8, p. 1 - 15.Gujarati, Damodar N., (1988), \"Basic Econometrics, Second Edition\", McGraw-Hill PublishingCompany, New York.Guy, D.C., John Hysom, and Stephen Ruth, (1985), \"The Effect of Subsidized Housing onValues of Adjacent Housing\", Journal of the American Real Estate and UrbanEconomics Association, 13.Haig, (1926), \"Toward an Understanding of the Metropolis: The Assignment of Activities toAreas in Urban Regions\", Quarterly Journal of Economics, 40.Hamilton, S.W., (1992), \"Residential Market Behavior: Turnover Rates and PricePerformance\", The Laurier Institute, Vancouver.Hamilton, S.W., and C. Carruthers, (1993), \"The Impacts of Hydro Transmission Lines onProperty Values\", The University of British Columbia, Vancouver.Henderson, J.V., (1980), \"Community Development: The Effects of Growth andUncertainty\", American Economic Review, 70, p. 894-910.63Henderson, J.V., (1985), \"The Impact of Zoning Policies Which Regulate Housing Quality\",Journal of Urban Economics, 18, p. 302-312.Jud, G., (1980), \"The Effects of Zoning on Single-family Residential Property Values:Charlotte, North Carolina\", Land Economics, 56, p. 142-154.Kain, J.K. and J.M. Quigley, (1970), \"Measuring the Value of Housing Quality\", Journal ofthe American Statistical Association, 65.Mark, J. and M. Goldberg, (1981), \"Land Use Controls: The Case of Zoning in the VancouverArea\", The American Real Estate and Urban Economics Association, 12, p. 418-435.Mark, J. and M. Goldberg, (1986), \"A Study of the Impacts of Zoning on Housing Valuesover Time\", Journal of Urban Economics, 20, p. 257-273.Michaels, G. and Kerry Smith, (1990), \"Market Segmentation and Valuing Amenities withHedonic Models: The Case of Hazardous Waste Sites\", Journal of Urban Economics, 28,p. 223-242.Mills, Edwin S. and Bruce W. Hamilton, (1989), Urban Economics 4th Edition, Glenview, Ill.:Scott, Foresman.Nourse, Hugh, (1963), \"The Effects of Public Housing on Property Values in St. Louis\",Land Economics, 39, p. 433-441.Palmquist, R. (1992), \"Valuing Localized Externalities\", Journal of Urban Economics, 31,p.59-68.Pearce, David W., (1989), The MIT Dictionary of Modern Economics, (Third Edition), TheMIT Press, Cambridge.Peterson, G. (1974), \"The Influence of Zoning Regulations on Land and Housing Prices\",Urban Institute Working Paper 1207-24.Pogodzinski, J. Michael and Tim R. Sass, (1990), \"The Economic Theory of Zoning: ACritical Review\", Land Economics, 66, p. 294-314.Rabiega, W.A. Ta-Win Lin, and Linda Robinson, (1984), \"The Property Value Impacts ofPublic Housing Projects in Low and Moderate Density Residential Neighborhoods\",Land Economics, 60, p. 174-179.Reuter, F.H., (1973), \"Externalities in Urban Property Markets: An Empirical Test of theZoning Ordinance in Pittsburg\", Journal of Law and Economics, 16, p. 315-350.64Rosen, S., (1974), \"Hedonic Prices and Implicit Markets: Product Differentiation in PureCompetition\", Journal of Political Economy, 82 (1), P. 34-55.Strange, William, (1992), \"Overlapping Neighborhoods and Housing Externalities\", Journal ofUrban Economics, 32, p. 17-39.Stull, W.J., (1974), \"Land-Use and Zoning in an Urban Economy\", American EconomicReview, 64, p. 337-347.Stull, W.J., (1975), \"Community Environment, Zoning, and the Market Value of SingleFamily Homes\", Journal of Law and Economics, 18, P. 535-557.Thibodeau, T.G. (1990), \"Estimating the Effect of High-Rise Office Buildings on ResidentialProperty Values\", Land Economics, 66, p. 402-408.Tunnicliffe, K., (1975), \"The Effect of Rezoning on Property Values: A Theoretical andEmpirical Examination of the Taxation of Land Value Increments Attributable toRezonings in the City of Vancouver\", M.Sc. thesis, Faculty of Commerce and BusinessAdministration, University of British Columbia.White, Michelle J., (1975), \"Effect of Zoning on the Size of the Metropolitan Area\", Journalof Urban Economics, 2, P. 279-290.Worchel, S., (1988), \"Understanding Social Psychology, Fourth Edition\", The Dorsey Press,Chicago.Young, G.M., (1943), \"Country and Town, A Summary of the Scott and Uthwatt Reports\",Penguin Books, Harmondsworth Middlesex, England.65"@en . "Thesis/Dissertation"@en . "1993-11"@en . "10.14288/1.0086205"@en . "eng"@en . "Business Administration"@en . "Vancouver : University of British Columbia Library"@en . "University of British Columbia"@en . "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en . "Graduate"@en . "The impacts of rezoning on adjacent property values"@en . "Text"@en . "http://hdl.handle.net/2429/2148"@en .