不同類型土壤及地下水污染場址外溢效果之評估──那種場址對房產價值減損最大？

Evaluating Different Types of Spillover Effects of Soil and Groundwater Contaminated Sites: Which Type Impair Housing Prices the Most?

土壤及地下水污染除了直接對受污染點的土地及房產價值產生影響外，透過消費者心理對場址厭惡心態之延伸，亦可能間接地對污染點外的鄰近土地及房產價值產生負面的外溢效果。此一議題迄今在國內尚未受到應有的重視及系統性的評估。有鑑於此，本文之目的即在評估不同類型土壤及地下水污染場址對鄰近房產之外溢效果。在實證上乃採用特徵價格法建立土壤及地下水污染場址與房產關聯之特徵價格函數，並以代理變數為外溢效果衡量的基礎，分別採用傳統迴歸模型與分量迴歸模型進行污染場址外溢效果的評估。結果顯示，加油站污染場址的房價外溢效果最高、亦即房產周邊有加油站對房產的減損最大，工廠類型次之，而農地類型最低，此一結果表示購屋者最不願與加油站類型的污染場址為鄰。此外，分量迴歸模型進而顯示，無論那一種類型污染場址，當房價越高、則外溢效果亦越高。在兩公里範圍內的顯著影響距離分群中，房產鄰近為加油站、農地或工廠等不同場址時，將使房價分別降低7.27％－51.54％、6.82％－14.81％及7.71％－36.20％。其次，若由距離分群的角度檢視，則無論在那一個區間下的房價，外溢效果對房價減損影響之百分比，在兩公里的範圍內相當一致的隨著距離的增加而減少。此一結果亦說明，擁有高區間房價的消費者，對於周圍是否存在污染場址相對於居住在低區間房價的消費者更加敏感。 The spillover impact of hazardous soil and groundwater contamination site on property value nearby the pollutant area is a high-profile research issue. The existing empirical results consistently show that the spillover effects on property value are significant as its distance from the contamination sites increase. However, there is still no systematic evaluation in Taiwan. Thus, the purpose of this study is to evaluate the spillover effects on property value caused by different types of contamination types of soil and groundwater pollution site. Empirically, the relationship connecting various attributes of housing property and its surrounding spatial attributes and characteristics of contamination sites is formulated in hedonic price method. A proxy variable is employed as an indicator of spillover effects of various kinds. The conventional OLS model and quantile regression model are adopted for comparison. The results show that housing property surrounded by contamination site of gas station impairs the housing price the most, by factory the send, and by farmland the least. That is, people are unwilling to pay to purchase housing property with gas station around. In addition, the delicate results can be observed from quantile regression that the higher the housing prices the larger spillover effects. Within 2 kilometer significant distance division, prices of housing properties surrounded by gas station, factory, and farmland will be decreased by 7.27%-51.54%, 6.82%-14.81%, and 7.71%-36.02% respectively. Furthermore, the results from quantile regression model show that in either price range of housing property the housing prices decrease consistently by either type of spillover of contamination sites. The results also show that people have high sensitivity for those who reside in housing properties with high prices than those reside in housing properties with low prices.