THE IMPACT OF AIRCRAFT NOISE ON HOUSE PRICES

S. Huderek-Glapska, R. Trojanek. The mpact of arcraft nose on house prces. Internatonal Journal of Academc Research Part B; 2013; 5(3), 397-408. DOI: 10.7813/2075-4124.2013/5-3/B.60 THE IMPACT OF AIRCRAFT NOISE ON HOUSE PRICES Sona Huderek-Glapska, Radosław Trojanek Poznań Unversty of Economcs (POLAND) E-mals: sona.huderek@ue.poznan.pl, r.trojanek@ue.poznan.pl DOI: 10.7813/2075-4124.2013/5-3/B.60 ABSTRACT The am of ths artcle s to dentfy the mpact of Warsaw Chopn Arport on house prces n Warsaw. Frederc Chopn Arport n Warsaw s the bggest arport and the man transfer node n Poland, whch servced 9.3 mllon passengers n 2011. Warsaw Chopn Arport s a cty arport, whch means that t s located wthn the borders of the cty of Warsaw. The vcnty of a large conurbaton makes the arport more accessble for people, although at the same tme ts operatons may cause some nconvenence to the local communty. One of the harmful effects of arport operaton s the nose and polluton emtted by planes usng the arport s nfrastructure. The am of ths study s to dentfy and measure the nfluence of Warsaw Chopn Arport on the prces of dwellngs located wthn the borders of the lmted use area. Ths research refers only to dwellngs located n mult-famly buldngs. Such choce was determned by two factors. Frstly, the majorty of dwellngs are located n mult-famly resdentals (dwellng blocks up to 90% n bg Polsh ctes). Secondly, houses are characterzed by great dfferentaton regardng both quanttatve and qualtatve features, whch requres the database to nclude the approprate nformaton about each property n order to construct house prce ndexes. In ths research the hedonc regresson method was used. Key words: Warsaw Chopn Arport, Frederc Chopn Arport, arcraft nose on house prces 1. INTRODUCTION Wth the growth of market economy n Poland, some markets whch played a margnal role n the pretranston economc system have grown n mportance. The housng market s undoubtedly one of them. The role ths segment of the real estate market plays n economy s determned by the fact that propertes are not only perceved as consumer goods, but also as captal, whch makes t possble to create added value for the owner as well as for local and natonal economy. A number of Polsh researchers have attempted to seek for relatonshps between dfferent markets of goods and the economy (1,2), effectveness of nvestments n dfferent segments of property market (3,4) or the mpact of globalzaton on whole markets (5). The mpact of arport operaton on house prces s one of the elements of the nfluence of an arport on ts surroundng area and belongs to external effects whch ndrectly affect the socety rather than drect users. Ths nfluence s reflected n the change of real estate prces caused by arport actvtes. These effects may be both postve (the ncrease of house prces) and negatve (the decrease of house prces). The areas surroundng an arport are usually well accessble and have good transport connectons wth the conurbaton. The road network around an arport s constantly developed and publc transport connectons (bus, tramway, underground, and ralway) are steadly mproved. Both the nhabtants of the area and the employees of companes based n the vcnty of an arport have easy and quck access to ar transport servces. Moreover, an arport wth ts modern servces adds to the attractve mage of the regon. All the above factors contrbute to the growth of demand for propertes located n the vcnty of an arport, whch n turn translates nto the ncrease of house prces. The negatve external effects are the result of the actvty of arlnes usng an arport s nfrastructure rather than the operatons of the arport tself. Arplanes emt nose and polluton, whch reduces the comfort of lvng among the populaton of the area around the arport. Moreover, there are numerous constrants concernng the use of propertes located near the arport. Those factors result n lower nterest n such propertes, reduce ther value and, consequently, contrbute to the drop n house prces. Baku, Azerbajan 397

Dagram 1. The nfluence of an arport on house prces A number of factors determne the drecton of an arport s nfluence on house prces. It s sometmes dffcult to clearly specfy the nature of an effect created by an arport. In some arport surroundng areas the external effects wll be postve, whle n others they wll be negatve. The ntensty of an arport s nfluence on house prces depends on: - the use of a property (housng, commercal, ndustral, publc), - the locaton of a property n relaton to an arport (how far from the arport t s located, whether t s located n the landng area), - the degree of an arport s nfluence, ncludng nose and polluton generaton (the number of ar operatons, the knd of arplanes used by arlnes). The nfluence of an arport on real estate prces s largely dependent on the purpose for whch a gven property s used. The demand for house prces shows the hghest senstvty to changes n the ntensty of an arport s nfluence caused by the ncrease of ar traffc. In the short term, the development of an arport causes negatve external effects for nhabtants due to the growth of the level of nose and ar polluton. As a result, the demand for dwellngs falls and ther prces decrease. In the long term, however, house prces may ncrease for two reasons. Frstly, thanks to replacng the nhabtants that are senstve to negatve externaltes of arport actvtes by those who are not senstve to such effects, but, qute the opposte, beleve that arport operaton benefts them. Secondly, prces rse when resdental propertes are transformed nto commercal ones. The knd of an arport s nfluence on the real estate market may be dentfed on the bass of the observaton of enttes locaton decsons (6). Changes n aggregate demand reflect choces made by market partcpants. If the demand for propertes located near an arport rses, we may conclude that postve externaltes preval over the negatve effects generated by arport operaton. Changes n real estate prces are determned by general and ndvdual factors. The general factors nfluencng the property value nclude, among other thngs, the ncrease of the populaton s dsposable ncome. The ndvdual factors result from the characterstcs of a gven knd of land (e.g. property locaton). In our analyss, we took nto account the ndvdual factors that are the result of the locaton of a property wthn the area nfluenced by arport operaton. 2. AN OVERVIEW OF STUDIES The dynamc development of the arlne ndustry and the ncrease n the number of ar operatons contrbuted to the ntensfcaton of fears about the deteroratng qualty of lfe of the people nhabtng the area affected by arport actvtes. The need has arsen for better understandng of the type and degree of an arport s nfluence upon ts surroundng, ncludng ts mpact upon the local communty, whch has to bear addtonal costs connected wth the ncreased nose and polluton emsson. Arspace s an example of a shared resource, whch could hardly be deemed a prvate property and whch s not subject to market operatons. Therefore, t s dffcult to evaluate ths resource and ts changes. Arlnes and ther consumers treat arspace as an nput n creatng a transport servce n the ar and they are wllng to pay for the rght to use ths space for ther own purposes (7). On the other hand, the local communty - whch nhabts the area surroundng the arport hghly values access to clean and peaceful envronment, thus, t s nclned to pay for t n ths specfc form. The externaltes of arport actvtes,.e. the emsson of nose and polluton, are not subject to market operatons and ther value s not objectvely assessed. There s a fear that those effects may be evaluated arbtrarly. State nterventon and the ntroducton of such regulatons that wll establsh the prncples of access to a gven resource offer an alternatve system for effectve resource allocaton. An example of such a regulaton could be the reducton of arcraft operatons at nght. 398 PART B. SOCIAL SCIENCES AND HUMANITIES

The externaltes resultng from arport operaton, partcularly arcraft nose, represent socal costs, whch could be evaluated usng the complementary market as a change n the value of propertes located n the area affected by arport actvtes. Potental buyers of these propertes are not always well nformed about the level of nose produced by arcraft operatons. The complete lack of or nadequate nformaton may lead to a hgh purchase prce and dsappontment after a purchase. After the level of nose stablzes, real estate prces adapt to new condtons and the next buyers buy propertes at dscounted prces, thus compensatng for the negatve effects resultng from the vcnty of an arport (6). The most frequently used methods of nose cost estmaton nclude: models based on revealed preferences and models based on stated preferences. Both approaches are based on the theory of consumer choce. Revealed preferences are consumers actual choces and they are analyzed wth the use of hstorcal data. Of all the models based on revealed preferences the hedonc prce model s the most frequently used method for analyzng the nfluence of arport operaton on house prces. Ths model s a knd of a parametrc method and t assumes that each good, n ths case a property, has a number of attrbutes whch have some contrbutory values consdered by a consumer. Ths s the most frequently used method for evaluatng the nfluence of publc projects, ncludng arports, upon the local envronment. Models based on revealed preferences make use of the exstence of complementary markets (the real estate market) to those on whch externaltes occur (the ar transport market). The dfference between the market value of propertes located n the area affected by arcraft nose and the value of those located beyond the arport surroundng area wth the other attrbutes unchanged represents the cost of nose. Ths dfference s usually expressed as nose dscount, whch decreases as the dstance from the arport ncreases. The analyss of the transacton regresson n the real estate market s used for dentfyng factors determnng the prce, thus estmatng the nfluence of arcraft nose. If the dfference n the market values of a property located wthn the arcraft nose zone and the one located outsde the area nfluenced by arcraft operaton s 40,000 PLN, and the dfference n the level of nose s 10dB, the dscount determned by the nose level s 4,000 PLN per 1 db wth the other attrbutes unchanged (8). In order to avod the use of nomnal values, the prce s expressed as the percentage of a property s value. Thus, f the market value of a property located n the nose zone s 360,000 PLN, and the value of the one located beyond that zone s 400,000 PLN, the NDI ndex (Nose Deprecaton Index) s 1.0% per db. The other knd of nose cost estmaton methods ncludes models based on stated preferences. They make use of economc surveys, are based on prmary data and enable the estmaton of non-market resources. The am of ths research s to analyze the ntentons consumers express when ntervewed. In ths method, we use the format of the questons about the wllngness to pay (WTP) for the mprovement of the present condton (nose reducton), whch wll result n the ncrease of usefulness. We also analyze the level of the acceptance of a negatve phenomenon (WTA wllngness to accept).one of the most popular methods based on stated preferences s the contngent valuaton (CV) method. It s a technque that apples the analyss of consumers preferences. Preferences are measured on the bass of the declaratons of preferences expressed n the approprate measurng scales. The surveyed consumers are presented the hypothetcal market for publc goods or servces and were offered a certan change (e.g. nose reducton by 10dB or reducng nose back to the level from before the extenson of an arport). Then, we analyze how much the consumers are wllng to pay for ths change (WTP). The external effect (nose) has as much value as the prce people affected by ths effect are ready to pay. In ths way the valuaton of nose costs s conducted. Methods based on stated preferences enable us to calculate the mnmum utlty of a gven good. Analyses based on WTP and WTA queston formats may be subjected to the econometrc analyss n order to estmate the mnmum cost of nose. Unlke methods based on revealed preferences, the ones based on stated preferences do not refer to market prces. Therefore, ther weakness les n ther relablty and nvarablty (n case the hypothetcal stuaton wll occur n realty). In comparson to parametrc methods, models based on stated preferences are relatvely new. However, the lterature provdes analyses of the nfluence of arcraft nose on real estate prces based on survey results. The frst research studes on the nfluence of arcraft nose upon the value of propertes located n the vcnty of an arport were carred out at the begnnng of the 1970s. They nvolved case studes of the arports n Toronto (6), Mnneapols (9), and London (10). Nelson (8) made an overvew of research done n the 1970s. On the bass of the results of 13 studes on the nfluence of arcraft nose upon the value of real estate, he ponted out that the nose deprecaton ndex (NDI) ranges from 0.4% to 1.1% per decbel. The dfference n the nose level was 20 db, and the propertes located n the nosest area were sold at 10-20% dscount. Later studes confrmed those fndngs. The study of the nfluence of the arport n Atlanta on the local envronment, ncludng the housng market (11) shows smlar dscount rates. However, the authors ndcate that earler studes from the 1960s and 70s were conducted durng the perod of changes n real estate markets, whch could affect research fndngs. Not all studes confrmed the lnk between the nose level and the property value. The study of the arport n Boston (12), carred out wth the applcaton of the hedonc prce model, showed that the relaton was statstcally rrelevant. By 2011 over 70 analyses of the nfluence of arcraft nose on the property value were performed. Most of them were conducted n the USA and Canada due to the fact that they have a more developed arlne market than other countres. The results of the analyses of the nfluence of arcraft nose on the property value dffer dependng on the number of arcraft operatons as well as on the sze and characterstcs of the area under study. Baku, Azerbajan 399

Table 1. The results of selected studes of the nfluence of arcraft nose on the value of propertes Author Arport Country Number of Results observatons (NDI) Emerson [1972] (9) Mnneapols USA 222 0.58 Dygert [1973] (12) San Francsco USA 128 0.50 Gautrn [1975] (13) London Unted Kngdom 67 0.62 De Vany [1976] (15) Dallas USA 1270 0.80 Maser n. [1977] (16) Rochester USA 1388 0.82-0.95 wthn cty borders 0.55-0.68 wthn conurbaton borders Nelson [1978] (8) Washngton USA 52 1.10 Meszkowsk, Saper [1978] (17) Toronto Canada 1200 0.52 Abelson [1979] (18) Sydney Australa 1414 0.40 Mc Mllan [1978] (19) Edmonton Canada 352 0.50 O'Byrne et al. [1985] (11) Atlanta Canada 258 0.52-0.67 Pennngton et al. [1990] (20) Manchester Unted Kngdom 3472 0.15 Uyeno et al. [1993] (21) Vancouver Canada 375 0.65-0.90 Levesque [1994] (22) Wnnpeg Canada 1635 1.30 Myles [1997] (23) Reno USA 4332 0.37 Tomkns et al. [1998] (24) Manchester Unted Kngdom 568 0.63 Salv [2003] (25) Zurych Swtzerland 565 0.75 McMllen [2004] (26) Chcago USA 4012 0.81 Baranzn, Ramrez [2005] (27) Geneva Swtzerland 1847 1.17 Cohen, Coughln [2008] (28) Atlanta USA 1643 0.43 The NDI ndex for Atlanta Arport s 0.43 (28), whch means that f the dfference n the level of nose between the zone nfluenced by arport operaton and the area beyond t, the propertes located n the arport surroundng area are sold wth a 4.3% dscount n relaton to smlar propertes located beyond that area, wth the other attrbutes unchanged. Dagram 2. The dstrbuton of the number of observatons accordng to the value of the NDI ndex On the bass of the studes of the nfluence of arcraft nose on real estate prces, researchers have carred out meta-analyses, thus generalzng the results obtaned earler (29-31). The nose deprecaton ndex (NDI) s defned as the percentage deprecaton of property values due to a unt ncrease n nose exposure. Nelson (8) carred out meta-analyss of 33 reported NDI estmates and concluded that the average NDI estmates n the USA and Canada was about 0.50% to 0.60% per db. The dfferences n NDI values were caused, among other thngs, by the type of the country s soco-economc profle, year of research, sze of study populaton, model specfcaton, and average real estate prces. Almost all analyses of the relaton between arcraft nose and real estate prces n the arport surroundng area showed the negatve nfluence of arport operaton on the value of propertes. The propertes nearest to the arport, n the landng area, were sold wth the bggest dscount. The further from the arport, the lower the dscount ndcator was. The knd of a property was mportant for the degree of the nfluence of nose on house prces. Hgh-standard houses and mult-famly resdentals were more vulnerable to the drop n prces. 400 PART B. SOCIAL SCIENCES AND HUMANITIES

3. RESEARCH METHODOLOGY AND SOURCES OF DATA In Polsh condtons, there s lmted access to data about transacton prces, especally hstorcal ones. 1 The data regardng dwellng prces orgnally covered over 400,000 offers of dwellngs (n Warsaw, durng the years 2007-2011) for sale. Both fully owned dwellngs and the ones wth a lmted rght of ownershp were examned. As a result of methodologcal selecton 2 the sze of the database was reduced to about 130,000 dwellngs for sale (n Warsaw. The number of offers gathered meets the requrements of the representatveness of a sample. Hedonc regresson Methods of the desgnng of house prce ndexes may be dvded, usng the crteron of allowng for changes n the quanttatve and qualtatve attrbutes of propertes, nto two groups: smple methods (those whch do not correct for such changes) and complex ones (those whch do allow for such changes, at least partly). Smple methods nclude methods based on the average and the ones based on the medan. Complex methods encompass: the hedonc regresson method, the resale method, the weghted average method and the hybrd one. In ths paper we shall focus on the hedonc regresson model. The frst documented use of the hedonc regresson dates back to 1922, when G. A. Hass developed the farmland prce model. As he publshed the results n the form of a techncal report, the real nfluence of ths research on the popularty of the hedonc method was far from sgnfcant (32). In 1926, Watt conducted a smlar study of farmland prces, whle n 1928 Waugh analyzed vegetable prces. However, t s Andrew Court who s consdered to be the father of the hedonc method. In 1939, he examned the nfluence of the attrbutes of cars on ther prces. The frst researcher to use the hedonc method to analyze the real estate market was probably Rdker he amed at dentfyng the nfluence of polluton reducton on house prces (33). The theoretcal framework of the hedonc method was developed by Lancaster (1966) and Rosen (1974). The essence of the hedonc method les n the assumpton that the prce of heterogeneous goods may be descrbed wth ts attrbutes. In other words, ths method may be used for estmatng the value of partcular attrbutes of a gven product. In order to dentfy the nfluence of ndvdual features on the value of a specfc good, econometrc equatons are constructed. The prce of a gven good s the response varable, whereas ts quanttatve and qualtatve attrbutes are the explanatory varables. The equaton may be recorded n the followng way: K P 0 X 1 u ; (1) where: P prce of a good β regresson coeffcent X attrbute of a good (value drver) u random error. The key ssue n hedonc methods s to choose the form of the regresson functon. The log-lnear form of the regresson functon s most frequently used for studyng changes n the prces n the real estate market n emprcal research: log P 0 K 1 X u ; (2) There are a few reasons for such a choce of functon(34). Frstly, the log-lnear model allows the added value (for example, the value resultng from the hgher standard) to change proportonally to changes of the sze and other attrbutes of the dwellng (n case of the lnear functon, for example, the mprovement of the standard wll have the same nfluence on the value of the dwellng wth the floor area of 30 m 2 and the one wth the surface area of 100 m 2, whereas n case of the log-lnear functon ths nfluence wll be dverse). Secondly, the estmated regresson coeffcents are easy to nterpret. The coeffcent of a gven varable may be defned as a percentage 1 For a few years a system of recordng prces and value has been developng at present, such data are avalable n most bg ctes. However, they only concern propertes (full ownershp), whle there s no nformaton about apartments wth a lmted rght of ownershp n mult-famly resdentals (such nformaton s only avalable n housng cooperatves n bg ctes there are several dozen of such nsttutons). There are dfferent ways of provdng these data n some ctes, e.g. n Wroclaw, they may be accessed onlne; n others, e.g. n Poznan, they are provded n the form of pdf prntouts (overall nformaton about transactons); whereas n Warsaw t s possble to browse and rewrte data only from the cards wth the most mportant nformaton ncluded n notaral deeds. The nformaton about dwellngs n mult-famly resdentals (wth a full or lmted rght of ownershp) ncluded n notaral deeds s ncomplete from the perspectve of ther applcablty for the constructon of ndexes wth the use of hedonc methods (lack of nformaton about the technology used, tme of constructon, techncal condton of the buldng, standard of decoraton). The stuaton s even worse n case of sngle-famly dwellng unts (houses). It s relatvely easy to obtan data about the date of transacton, the locaton, the sze of a plot, whereas t s hardly possble to fnd nformaton about the constructon technology, the tme of constructon, the techncal condton, the nteror decoraton standard, the usable floor space. Moreover, even f t s provded, ts type s not specfed whether t s the total usable area; f t s usable, what norm was used for estmatng t; the floor space for tax calculatons ). 2 Empty and recurrng records were removed as well as those n whch a specfc offer was not fully descrbed. The recurrence of data was the result of announcng one offer by a few estate agents, thus they were repeatedly placed n a database. The next stage of the analyss nvolved checkng the relablty of the obtaned data. The am was to elmnate those offers whch were, for no clearly specfed reasons, far from the average. Moreover, t was assumed that the analyss wll cover apartments wth the floor space of up to 150 m 2 and havng no more than fve rooms. Both fully owned apartments and the ones wth a lmted rght of ownershp were examned. Baku, Azerbajan 401

change of the value of an dwellng caused by the unt change of a value drver. Thrdly, the log-lnear functon often eases problems connected wth heteroscedastcty or wth the varablty of a random component. The hedonc method has a lot of applcatons n the research of the real estate market, the most mportant of whch seems to be ts applcablty n the constructon of real estate prce ndexes. Indexes of house prces based on the hedonc regresson may be bult n two man ways (35)]: - on the bass of the equatons of dwellng prces constructed for each of the perods under analyss or - on the bass of one equaton of dwellng prces constructed for two or more perods. In the frst approach, regresson models of house prces for each perod under study are desgned. The value of each property at a gven tme s dfferent on account on ther qualtatve (such as type of housng or locaton) and quanttatve (for example, number of rooms, number of bathrooms, age of the buldng) features. The value of each property may be represented as the functon of ther measurable attrbutes X 1 and mmeasurable attrbutes u, whch are specfc to each property, but we do not have access to any data concernng them. Ths relatonshp may be represented by the followng equaton: log( p ) 0 j X u 1 j j ; (3) where: p prce of a property, β regresson coeffcent, X attrbute of a good (value drver), u random error. Wthn ths approach, we may dstngush the characterstcs prce method and the mputaton methods. In the characterstcs prce method, the average values of dwellng features n a selected perod are establshed n ths way the states of the attrbutes of an average dwellng are specfed. Then, usng the estmated regresson coeffcents n varous perods and the specfed states of the attrbutes of an average dwellng, the prce of an average dwellng s estmated for each perod. They serve as the bass for the constructon of an ndex. In the mputaton method, the estmated econometrc equatons for dfferent perods are used for establshng the value of dwellngs from the base perod. In other words, the states of the attrbutes of dwellngs from the reference perod are placed n the regresson equatons estmated for the analyzed perods. Thus, the value of a fxed basket of dwellngs wth the same state of attrbutes n dfferent perods s establshed. In the other approach, the house prce regresson equaton s constructed. It ncludes the tme dummy varable. where: log( p ) 0 T j Q 1 t 1 X j j u ; (4) Q tme dummy varable (t takes the value 1 f a gven observaton was made n perod τ; otherwse t takes zero). Wthn ths approach, we may dstngush two more varants the regresson equaton s constructed for two neghbourng perods (adjacent perod tme dummy varable method) and for more than two neghbourng perods (pooled tme dummy varable method). The prncpal dfference between the two approaches les n the fact that n the frst case both the average and standard devaton of random error dffers dependng on the perod under study, whle n the other they are constant. The adopton of the hedonc method for studyng changes n the market for dwellngs requres a lot of effort whle gatherng data, because not only do we need the nformaton about real estate prces ndspensable, but we also have to fnd out about the states of the attrbutes of each property. If there s no suffcent database, whch wll nclude all the necessary nformaton concernng the states of real estate attrbutes, the hedonc method may not provde a relable house prce ndcator n a specfc perod. 4. CHANGES IN THE PRICES OF DWELLINGS IN WARSAW AND DISTRICTS WITHIN THE LIMITED USE AREA IN 2007-2011 In order to dentfy house prce ndexes wth the use of the hedonc regresson method, the nformaton on askng prces for resdental propertes n Warsaw n the years 2007-2011was collected (36). In the study, we used the hedonc method based on the regresson equaton of the prces of dwellngs, ncludng the tme dummy varable (4). The choce of qualtatve and quanttatve data was lmted by the avalablty of nformaton n the database. Table 1 presents varables used n the study. 402 PART B. SOCIAL SCIENCES AND HUMANITIES

Table 2. Qualtatve and quanttatve varables appled n the model Varable Symbol Descrpton Perod Q1 2007 1 q. 20 tme dummy varables. It takes the value 1 f the dwellng was. offered for sale n a gven perod; otherwse, t takes 0. Q20 0 2011 4 q. Locaton d1. d18 Tme dummy varables. The admnstratve dvson of cty was used. If the dwellng s located n a gven dstrct, t takes the value 1; otherwse t takes 0). Materal m 1-prefabrcated 2-tradtonal technology Tme of constructon R1 before 1939 R2 from 1945 to 1989 R3 from 1945 to 2000 R4- after 2000 4 tme dummy varables. If the dwellng s placed n a buldng bult n a gven perod, t takes the value 1; otherwse t takes 0). Floor space pow The floor area of a gven dwellng s measured n square metres. Floor p 1 ground and last floor 2 - ntermedate floors 3 - frst and second floor Type of ownershp pw 1 full ownershp 2 lmted rght of ownershp Standard s It takes value 1 for dwellngs wth the lowest standard, and 5 for those wth the hghest. Arcraft nose ST1 It takes value 1 for dwellngs located n the lmted use area (LUA) Then, usng GRETL software, the econometrc equatons for each quarter of the perod between 2007 and 2011 were estmated n the form of the above equaton (4), n whch the response varables ncluded the locaton, constructon materal, standard, type of ownershp, tme of constructon and floor space of an dwellng. Table 3 presents the results of the house prce regresson functon for the dstrcts of Warsaw n 2007-2011. Table 3. The results of the house prce regresson functon for the dstrcts of Warsaw n 2007-2011 Bemowo Mokotów Ochota Ursynów Wola const 12.0928 12.1225 12.101 12.2885 12.1179 q1 0.06519 0.136246 0.092588 0.069078 0.073578 q2 0.07752 0.149023 0.136623 0.091515 0.098099 q3 0.12475 0.133039 0.125406 0.085646 0.096181 q4 0.08676 0.119387 0.132854 0.07514 0.086085 q5 0.08815 0.116518 0.122031 0.081334 0.086514 q6 0.07244 0.115173 0.074551 0.070158 0.084297 q7 0.05057 0.107527 0.085107 0.06358 0.041496 q8 0.04989 0.093268 0.061044 0.05501 0.025126 q9 0.00835 0.062151 0.034525 0.011034-0.00563 q10 0.01289 0.055086 0.028955 0.010725-0.00843 q11 0.01185 0.032373 0.026827 0.018337-0.00228 q12 0.01645 0.061157 0.046097 0.030457-0.00363 q13 0.01669 0.058595 0.062512 0.03695 0.026745 q14 0.03383 0.072738 0.063978 0.043384 0.040566 q15 0.02309 0.09542 0.056823 0.046081 0.038228 q16 0.03182 0.063567 0.036133 0.033973 0.017499 q17 0.04209 0.060418 0.062102 0.053504 0.038301 q18 0.02557 0.046215 0.022097 0.031986 0.048894 q19 0.00214 0.015295-0.01012 0.016157-0.00291 floor space 0.01421 0.015947 0.015831 0.012518 0.016336 floor 0.00953 0.013749 0.008116 0.022627 0.009147 m 0.03475 0.086391 0.054435 0.050195 0.020655 pw -0.0161-0.03804-0.01966-0.01234 0.000423 s 0.03264 0.023124 0.02844 0.023892 0.020473 R1-0.0146-0.08238-0.05041-0.08971-0.20253 R2-0.1698-0.18878-0.18414-0.17482-0.19671 R3-0.1199-0.16236-0.09097-0.07644-0.11682 R-squared coeffcent of determnaton 0.85 0.87 0.89 0.88 0.88 Baku, Azerbajan 403

On the bass of the obtaned results t may be ponted out that the explanatory varables used n the equaton to a large extent (n almost 90%) explan the fluctuatons of the prces of dwellngs n Warsaw n 2007-2011. Moreover, all varables used n the model turned out to be statstcally relevant. On the bass of the above estmatons we establshed dwellng prce ndexes for the selected dstrcts and the cty of Warsaw n 2007-2011 (1 st quarter of 2007 = 100). Dagram 5 and table 3 present dwellng prce ndexes desgned on the bass of the hedonc regresson n the selected dstrcts of Warsaw and n Warsaw n 2007-2011. Dagram 3. Dwellng prce ndexes n the selected dstrcts of Warsaw and n Warsaw n 2007-2011 (1 st quarter of 2007 = 100) The ndexes presented n the dagram and the table fluctuated n a smlar way, but they dffered n the ampltude of prce fluctuatons n the partcular dstrcts. The ndex of the nomnal prces of dwellngs n Warsaw (1 st quarter of 2007 = 100) n the 4 th quarter of 2011 was 91.07. It means that the nomnal prces of dwellngs n Warsaw dropped by about 9%. Prce drops n the partcular dstrcts were: Mokotów 12%, Ochota 8.5%, Bemowo 6%, Wola 7%, Ursynów 6.5%. Comparng ndexes of dwellng prces n Warsaw and the analyzed dstrcts n the analyzed perod, we may observe that some dstrcts were more resstant to drops n prces. In the perod under study, the dstrct of Mokotów wtnessed the bggest drops, whle n the dstrct of Bemowo prces fell the least. Another dagram shows the prce ndex of the dwellngs located n the lmted use area (LUA) and the ones located beyond t n Warsaw n 2007-2011 (regresson functon estmatons are presented n the appendx - tables 11, 12). Dagram 4. The prce ndexes of the dwellngs located n the area nfluenced of arport operaton and the ones located beyond t n Warsaw n 2007-2011 (1 st quarter of 2007 = 100) 404 PART B. SOCIAL SCIENCES AND HUMANITIES

The ndexes presented n the dagram fluctuated n a smlar way and dffered slghtly n the ampltude of prce fluctuatons n partcular perods. The analyss of the dagram shows that n the perod under study the prces of the dwellngs located outsde the LUA fell by 0.5% more than the prces of dwellngs wthn the LUA. 5. THE IDENTIFICATION OF THE INFLUENCE OF THE LUA ON THE PRICES OF DWELLINGS IN WARSAW WITH THE APPLICATION OF THE HEDONIC REGRESSION In order to establsh the nfluence of the LUA on the prces of dwellngs, we used the hedonc regresson method. In the study, we used askng prces for resdental propertes n Warsaw n the years 2007-2011. The choce of qualtatve and quanttatve data was lmted by the avalablty of nformaton n the database. Table 4 presents varables used n the study. Then, usng GRETL software, we estmated the parameters of functons, n whch the prce of an dwellng was the response varable, whle the explanatory varables ncluded the locaton, constructon materal, standard, type of ownershp, tme of constructon, floor space, number of rooms and the locaton n the LUA. Table 5 presents the results of the regresson functon for the equaton. Table 4. The estmates of prce functon parameters, used observatons 1-130324, dependent varable: prce log. Coeffcent Standard error Student's t- dstrbuton p value const 12.1748 0.00458138 2657.4457 <0.00001 *** d1-0.140675 0.00331367-42.4528 <0.00001 *** d2-0.360044 0.00327917-109.7973 <0.00001 *** d3-0.100006 0.00318609-31.3883 <0.00001 *** d4 0.0151945 0.00288138 5.2733 <0.00001 *** d5-0.0291474 0.00322338-9.0425 <0.00001 *** d6-0.136572 0.00303072-45.0627 <0.00001 *** d7-0.209871 0.0038255-54.8611 <0.00001 *** d8-0.338431 0.00696623-48.5817 <0.00001 *** d9 0.158795 0.00296089 53.6307 <0.00001 *** d10-0.243402 0.0034762-70.0195 <0.00001 *** d11-0.249484 0.0038278-65.1769 <0.00001 *** d12-0.0427418 0.00336554-12.6998 <0.00001 *** d13-0.300017 0.00499708-60.0385 <0.00001 *** d14-0.34247 0.00782507-43.7657 <0.00001 *** d15-0.133747 0.00384812-34.7565 <0.00001 *** d16-0.201503 0.00447062-45.0728 <0.00001 *** d17-0.0778461 0.00314341-24.7649 <0.00001 *** q1 0.0980847 0.006132 15.9955 <0.00001 *** q2 0.113533 0.0048282 23.5146 <0.00001 *** q3 0.113908 0.00402818 28.2778 <0.00001 *** q4 0.107341 0.00339174 31.6477 <0.00001 *** q5 0.105063 0.00321149 32.7146 <0.00001 *** q6 0.0939037 0.0029962 31.3409 <0.00001 *** q7 0.0783384 0.00271463 28.8578 <0.00001 *** q8 0.06549 0.00327267 20.0112 <0.00001 *** q9 0.0278972 0.00282307 9.8819 <0.00001 *** q10 0.0308345 0.00266002 11.5918 <0.00001 *** q11 0.0238384 0.00325355 7.3269 <0.00001 *** q12 0.041398 0.00301622 13.7251 <0.00001 *** q13 0.0518851 0.00268926 19.2935 <0.00001 *** q14 0.0578567 0.00274854 21.0500 <0.00001 *** q15 0.0586741 0.00291818 20.1064 <0.00001 *** q16 0.0414393 0.00280327 14.7825 <0.00001 *** q17 0.0507324 0.00269629 18.8156 <0.00001 *** Baku, Azerbajan 405

q18 0.035035 0.00255473 13.7138 <0.00001 *** q19 0.0116182 0.00274885 4.2266 0.00002 *** floor space 0.0151588 2.22401e-05 681.6013 <0.00001 *** floor 0.012109 0.000701154 17.2702 <0.00001 *** m 0.0544772 0.00151655 35.9217 <0.00001 *** pw -0.0168125 0.00120454-13.9575 <0.00001 *** s 0.0243807 0.000326736 74.6190 <0.00001 *** R1-0.0652331 0.00197821-32.9758 <0.00001 *** R2-0.173228 0.00151154-114.6039 <0.00001 *** R3-0.101224 0.00161233-62.7814 <0.00001 *** ST1-0.0105966 0.0017852-5.9358 <0.00001 *** Mean of Dependent Varable 13.09661 Std. Dev. of Dependent Varable 0.462831 Resdual sum of squares 3748.417 Standard Error of the Resduals 0.169683 Coeffcent of Determnaton R 2 0.865637 Adjusted R 2 0.865590 On the bass of the obtaned results t may be concluded that the explanatory varables used n the equaton explan the fluctuatons of dwellng prces n Warsaw n 2007-2011 n 87%. Moreover, all the varables appled n the model turned out to be statstcally relevant. From the analytcal pont of vew, the statstcal relevance of ST1 varable s extremely mportant. The applcaton of the log-lnear model let to dentfy the percentage dfference n the prce of the same dwellng located wthn the LUA and outsde ths zone. In our case, the value of the coeffcent wth ST1 varable s -0.01, whch ndcates that an dwellng located n the LUA was about 1% cheaper than the same dwellng located beyond ths area n Warsaw n the years 2007-2011. Thus, we may conclude that the locaton of an dwellng wthn the LUA reduces ts value by 1% on average. When compared to the mpact of other varables on the prce, ths nfluence s margnal. The research results show a hgh analytcal value of askng prces, especally n the cases of the lmted access to and small number of transacton prces n a research sample meetng the requrements of representatveness As the collecton of avalable askng prces n the analyzed perod of 2010-2011 s very large (130,000 tems), we were also able to estmate prce equatons for the partcular dstrcts of Warsaw located wthn the LUA. the mpact of the LUA upon the askng prces of dwellngs n the dstrcts s shown n Dagram 5, whch presents dfferences between prces of dwellngs located wthn the LUA ad those outsde the zone. Dagram 5. Dfferences between prces of dwellngs located n the LUA and those beyond ths area The analyss of the results concernng prces for the whole set of resdental propertes n Warsaw shows that the prce of an dwellng located wthn the LUA s 1% lower than the prce of an dwellng located beyond t. As regards dwellngs n the partcular dstrcts, the negatve nfluence of the locaton n the LUA on dwellng prces (askng prces) s dverse. In the case of such dstrcts as Włochy or Ursynów, the locaton wthn the LUA dd not affect dwellng prces (coeffcents turned out to be rrelevant). In the case of the dstrcts of Bemowo, Mokotów, Ochota and Wola, the mpact of the locaton wthn the LUA turned out to be negatve the hghest degree of nfluence (4.9%) n the case of dwellngs n the dstrct of Wola, 2 % for Mokotów, 1.9 % for Bemowo and 1 % for dwellngs n Ochota. 406 PART B. SOCIAL SCIENCES AND HUMANITIES

6. CONCLUSION The am of ths artcle was to dentfy the mpact of arcraft nose created by Warsaw Chopn Arport on dwellng prces n Warsaw. Hedonc regresson method was used to estmate the mpact of arcraft nose on dwellng prces. The applcaton of the log-lnear model let to dentfy the percentage dfference n the prce of the same dwellng located wthn the LUA and outsde ths zone. In case of ths research, the value of the coeffcent wth ST1 varable s -0.01, whch ndcates that an dwellng located n the LUA was about 1% cheaper than the same dwellng located beyond ths area n Warsaw n the years 2007-2011. As regards dwellngs n the partcular dstrcts, the negatve nfluence of the locaton n the LUA on dwellng prces (askng prces) s dverse. In the case of such dstrcts as Włochy or Ursynów, the locaton wthn the LUA dd not affect dwellng prces (coeffcents turned out to be rrelevant). In the case of the dstrcts of Bemowo, Mokotów, Ochota and Wola, the mpact of the locaton wthn the LUA turned out to be negatve the hghest degree of nfluence (4.9%) n the case of dwellngs n the dstrct of Wola, 2 % for Mokotów, 1.9 % for Bemowo and 1 % for dwellngs n Ochota. REFERENCES 1. R. Kasperowcz, Forecastng the consumpton of electrcty by Polsh ndustry, Actual Problems of Economcs, #2, pp. 81-90, (2013) 2. R. Kasperowcz, The Co-occurrence of Cyclcal Fluctuatons of Economc Aggregates n Polsh Manufacturng Industry, Actual Problems of Economcs, vol.2 #7, pp. 63-72, (2012) 3. L. Nawrot, Bomass energy nvestment n rural areas possble applcatons n toursm ndustry, Actual Problems of Economcs, Vol. 2, #7, 140-150, (2012). 4. M. Bednarska, L. Nawrot, Determnants of lodgng ndustry valuaton for assessment purposes, Actual Problems of Economcs, Vol. 2, #7, 24-32, (2012). 5. K. A. Nawrot, Current challenges to the changng pattern of East Asan Cooperaton and Governance European Perspectve, n: Actual Problems of Economcs Vol. 2, #7, pp. 15-23, (2012). 6. R.W. Crowley. A case study of the Effects of an Arport on Land Values. Journal of Transport Economcs and Polcy 7: 144-152 (1973). 7. D. Gllen. The economcs of nose. (n) D. A. Hensher; K.J. Button (ed.) Handbook of Transport and the Envronment. Elsever. Amsterdam. 81-95 (2003) 8. J. P. Nelson. Arports and Property Values: a Survey of Recent Evdence, Journal of Transport Economcs and Polcy 14, 37-52 (1980). 9. F.C. Emerson. Valuaton of resdental amentes: an econometrc approach. Apprasal Journal 40: 268-278 (1972). 10. A. A. Walters. Nose and Prces. Oxford Unversty Press. London (1975). 11. N. P. H. O Byrne; J. J. Seneca. Housng values, census estmates, dsequlbrum, and the envronmental costs of arport nose: A case study of Atlanta. Journal of Envronmental Economcs And Management, 12: 169-178 (1985) 12. M. L, H. J. Brown. Mcro-neghborhood externaltes and hedonc housng prces. Land Economcs 56: 125-141. (1980). 13. P. K. Dygert. Estmaton of the Cost of Arcraft Nose to Resdental Actvtes. Unversty of Mchgan, Ph. D. dssertaton. (1973). 14. J. F. Gautrn. An evaluaton of the mpact of arcraft nose on property values wth a smple model of urban land rent, Land Economcs 51: 80-86 (1975). 15. A. De Vany; J. P. Nelson, A. A. Wlaters. Cost-beneft analyss: some llustratons, (w:) Nose Abatement Polcy Alternatves for Transportaton, Washngton: 205-214 (1977). 16. S. M. Maser; W. H. Rker, R. N. Rosett. The effects of Zonng and Externaltes on the Prce of Land: An Emprcal Analyss of Monroe County, New York, Journal of Law and Economcs, 20: 111-132 (1977). 17. P. Meszkowsk; A. M. Saper. An Estmate of the Effects of Arport Nose on Property Values: 425-440 (1978). 18. P. W. Abelson. Property Prces and the Value of Amentes, Journal of Envronmental Economcs and Management, 6: 11-28 (1979). 19. M. L. McMllan; B. G. Red; D. W. Gllen. An Approach towards Improved Estmates of Wllngness to Pay for Publc Goods from Hedonc Prce Functons: A case of Arcraft Nose, Unversty of Alberta: 315-328 (1978). 20. G. Pennngton; N. Tophan and R. Ward. Arcrafts nose and resdental property values adjacent to Manchester Internatonal Arport, Journal of Transport Economcs and Polcy, 24: 49-59 (1990). 21. D. Uyeno; S. W. Hamlton; A. Bggs. Densty of resdental land use and the mpact of arport nose, Journal of Transport Economcs and Polcy, 27: 3-18 (1993). 22. Levesque T., Modelng the effects of arport nose on resdental housng markets: A case study of Wnnpeg Internatonal Arport, Journal of Transport Economcs and Polcy, 28: 199-210 (1994). 23. C. L. Myles. Sound Effects: Measurng the Impact of Arcraft Nose on Resdental Property Values Around McCarran Internatonal Arport, unpublshed M.A. thess, Unversty of Nevada, Las Vegas (1997). 24. J. Tomkns; N. Topham; J. Twomey; R. Ward. Nose vs. Access: The Impact of an Arport n an Urban Property Market. Urban Studes. 35: 243-258 (1998). Baku, Azerbajan 407

25. M. Salve. Spatal Estmaton of the Impact of Arport Nose on Resdental Housng Prces, Workng Paper (2003). 26. D. P. McMllen. Arport expansons and property values: The case of Chcago O Hare Arport, Journal of Urban Economcs, 55: 627-640 (2004). 27. A. Baranzn and J. V. Ramrez. Payng for Quetness: the Impact of Nose on Geneva Rents. Urban studes, 42(4): 633-646 (2005). 28. J. P. Cohen and C. C. Coughln. Spatal Hedonc Models of Arport Nose, Proxmty, and Housng Prces, Journal of Regonal Scence, Wley Blackwell, vol. 48(5): 859-878 (2008). 29. Y. Shpper, P. Njkamp, P. Retveld. Why Do Arcraft Nose Value Estmates Dffer? A Meta-Analyss, Journal of Ar Transport Management, 4: 117-124 (1998). 30. I. Batema;, B. Day, I. Lake; A.. Lovett.The Effect of Road Traffc Nose on Resdental Property Values: A Lterature Revew and Hedonc Prcng Study. Ednburgh: Scottsh Executve Development Department (2001) 31. J. Nellthrop A. L. Brstow, B. Day. Introducng wllngness-to-pay for nose changes nto transport apprasals: an applcaton of beneft transfer, Transport Revews 27: 327-353 (2007). 32. P.Colwell, Dlmore, G., Who was Frst: An Examnaton of an Early Hedonc Study. Land Economcs, vol. 75. (1999) 33. E. Coulson, Monograph on Hedonc Estmaton and Housng Markets, Department of Economcs, Penn State Unversty, pp. 3-76, (2008) 34. S. Malpezz, Hedonc Prcng Models: A Selectve and Appled Revew, w Housng Economcs and Publc Polcy: Essays n honor of Duncan Maclennan, red. T. O Sullvan, K. Gbb, Oxford: Blackwell, pp. 67-89, (2008) 35. S.C Bourassa, Hoesl, M., Sun, J., 2006, A Smple Alternatve House Prce Index Method, Journal of Housng Economcs, vol. 15, pp. 80-97, (2006). 36. R. Trojanek, An analyss of changes n dwellng prces n the bggest ctes of Poland n 2008-2012 conducted wth the applcaton of the hedonc method, Actual Problems of Economcs 2012 # 7, pp. 5-14 (2012). 408 PART B. SOCIAL SCIENCES AND HUMANITIES

Copyrght of Internatonal Journal of Academc Research s the property of Internatonal Journal of Academc Research and ts content may not be coped or emaled to multple stes or posted to a lstserv wthout the copyrght holder's express wrtten permsson. However, users may prnt, download, or emal artcles for ndvdual use.

Copyrght of Internatonal Journal of Academc Research s the property of Internatonal Journal of Academc Research and ts content may not be coped or emaled to multple stes or posted to a lstserv wthout the copyrght holder's express wrtten permsson. However, users may prnt, download, or emal artcles for ndvdual use.