# RECORD 85ec88163a881d2f9a23ced6110d5e4e05b48d09 # TITLE: Urban Traffic Arterials and the Sale Price of Single Family Houses in Portland: A Hedonic Analysis of Impacts # YEAR: 2022.0 VENUE: nan DOI: 10.15760/etd.8061 # AUTHORS: P. Wuest # PAGES: 84 CANDIDATE TABLE PAGES (1-based): [28, 46, 47, 49] # ABSTRACT: An abstract of the thesis of Philip Wuest for the Master of Urban Studies degree presented August 6, 1997. Title: Urban Traffic Arterials and the Sale Price of Single Family Houses in Portland: A Hedonic Analysis of Impacts. This thesis explores the hypothesis that the prices of single family houses are negatively influenced by proximity to urban arterial roadways. A hedonic price.model is developed that tests the effects of traffic volumes and proximity on sale price. Using a data set of sale records from Portland Oregon in 1994, a traditional hedonic model is developed based on housing and neighborhood characteristics. Using a geographic information system, each sale record is associated with the nearest arterial roadway, and the data on distance to the roadway and level of traffic is returned to the sale record. These data are then applied in the hedonic equation. The results indicate that distance to an arterial roadway has a much larger effect than traffic volume alone. It is also found that the effects of high traffic arterials diminish rapidly with distance, becoming difficult to distinguish beyond about 2/3 of a city block. The estimated effect on sale price of close proximity to a busy roadway is large, approximately -10% of sale price. 2 URBAN TRAFFIC ARTERIALS AND THE SALE PRICE OF SINGLE FAMILY HOUSES IN PORTLAND: A HEDONIC ANALYSIS OF IMPACTS ===== PAGE 1 (score 4, numeric-paren 0, table-words 0, var-words 4) ===== Portland State University Portland State University PDXScholar PDXScholar Dissertations and Theses Dissertations and Theses 8-6-1997 Urban Traffic Arterials and the Sale Price of Single Urban Traffic Arterials and the Sale Price of Single Family Houses in Portland: A Hedonic Analysis of Family Houses in Portland: A Hedonic Analysis of Impacts Impacts Philip Wuest Portland State University Follow this and additional works at: https://pdxscholar.library.pdx.edu/open_access_etds Part of the Urban Studies and Planning Commons Let us know how access to this document benefits you. Recommended Citation Recommended Citation Wuest, Philip, "Urban Traffic Arterials and the Sale Price of Single Family Houses in Portland: A Hedonic Analysis of Impacts" (1997). Dissertations and Theses. Paper 6201. https://doi.org/10.15760/etd.8061 This Thesis is brought to you for free and open access. It has been accepted for inclusion in Dissertations and Theses by an authorized administrator of PDXScholar. Please contact us if we can make this document more accessible: pdxscholar@pdx.edu. ===== PAGE 28 (score 20, numeric-paren 0, table-words 2, var-words 5) ===== Table 2: Descriptive Statistics Name Variable N Mean St.Dev. Minimum Maximum LN PRC Y2 1437 11.33 0.44 10.31 12.89 BATHRMS X1 1437 1.25 0.51 0.00 5.00 FIREPLC X3 1437 0.79 0.66 0.00 3.00 GARAGE X4 1437 1.00 0.74 0.00 5.00 SQFTBLD XS 1437 1292.00 493.48 400.00 4032.00 SQFTBD2 X6 1437 1912700.00 1691400.00 160000.00 16257000.00 SQFTLOT X7 1437 6767.00 3451.80 1850.00 31800.00 SQFTLOT2 XS 1437 57699000.00 80476000.00 3422500.00 1011200000.00 STORIES X9 1437 1.10 0.30 0.00 2.00 VIEW X10 1437 0.07 0.53 0.00 6.00 STRUCAGE X11 1437 55.15 24.10 0.00 134.00 STRUCAG2 X12 1437 3621.50 2545.80 0.00 17956.00 NEGHAGE XX4 1437 10.43 9.53 0.34 65.91 VANDAL XX6 1437 16.03 6.83 2.00 65.00 BURGLAR xxs 1437 3.42 1.45 1.00 8.00 MDFAMINC XX10 1437 31800.00 9547.80 14625.00 82920.00 NONWHTE XX11 1437 16.99 15.04 0.00 78.06 PCTBADEG XX12 1437 13.62 8.54 2.29 44.99 HIGH85 V12 1437 0.02 0.15 0.00 1.00 HIGH350 V13 1437 0.20 0.40 0.00 1.00 HIGH550 V14 1437 0.15 0.35 0.00 1.00 Figure 1: Frequency Distribution: sale price and ln(sale price) HIST0'.3RAM - Sale Price PCT. N 0.602 865 IXXXXXXXXXX 0.577 829 IXXXXXXXXXX 0.552 793 IXXXXXXXXXX 0.527 757 IXXXXXXXXXX 0.502 721 IXXXXXXXXXX 0.477 685 IXXXXXXXXXX 0.452 649 IXXXXXXXXXX 0 427 613 IXXXXXXXXXX 0.402 577 IXXXXXXXXXX 0.376 541 IXXXXXXXXXX G.351 505 IXXXXXXXXXX 0.326 469 IXXXXXXXXXX 0.301 433 IXXXXXXXXXX 0.276 397 IXXXXXXXXXXXXXXX 0.251 361 IXXXXXXXXXXXXXXX 0.226 325 IXXXXXXXXXXXXXXX 0 201 289 IXXXXXXXXXXXXXXX 0.176 253 IXXXXXXXXXXXXXXX 0.151 217 IXXXXXXXXXXXXXXX 0.126 181 IXXXXXXXXXXXXXXX 0.101 145 IXXXXXXXXXXXXXXX 0.076 109 IXXXXXXXXXXXXXXX 0.051 73 I 0.026 37 I 0.001 1 I I---------I---------I---------I---------I---------1---------I 0.300E+05 0.908E+05 0.152E+06 0.213E+06 0.273E+06 0.334E+06 0.395E+06 19 ===== PAGE 29 (score 3, numeric-paren 0, table-words 1, var-words 1) ===== HIS'I'CGRA.M -ln(sale price) PCT, N 0,351 505 0,337 484 0,322 463 0,308 442 0,293 421 0,278 400 0,264 379 0.249 358 0,235 337 0 220 316 0,205 295 0,191 274 0,176 253 0,161 232 0,147 211 0,132 190 0,118 169 0,103 148 0,088 127 0,074 106 0,059 85 0,045 64 0,030 43 0,015 22 0,001 1 I I I I I I I I I I I I I I I I I I I I I I I I I xxxxxxxxxx xxxxxxxxxx xxxxxxxxxx xxxxxxxxxx xxxxxxxxxx xxxxxxxxxx xxxxxxxxxx xxxxxxxxxxxxxxx xxxxxxxxxxxxxxx xxxxxxx:xxxxxxx I---------I---------1---------I---------I---------I---------l 10,3 10,7 11,2 11,6 12,0 12,5 12,9 Development of the Arterial Variables: The Role of GIS Geographic information systems are a powerful tool in conducting spatial analysis of urban structure, attributes and systems. This section will describe the process that developed several of the components of the variables used in this study. The beginning data set for this analysis contained 13,589 records; a complete record of all sales of single family homes in the City of Portland for 1994 (Figure 2). Each sale record was connected to both parcel attributes from the Multnomah County tax assessor and were further enhanced with neighborhood and structure attributes from 20 ===== PAGE 46 (score 32, numeric-paren 0, table-words 6, var-words 5) ===== w ---.J Table 5: C ftest model Dependent variable: natural log of sale price Model 1: Adj. R"2 = .595 Var. Name Coefficients t-score x1 BATHRMS 0.0439 1.705 x3 FIREPLC 0.0775 5.38 x4 GARAGE 0.0379 2.638 x5 SQFTBLD 0.0006 8.467 x6 SQFTBD2 -0.0000001 -4.341 x7 SQFTLOT 0.00001 1.23 x8 SQFTLOT2 -1E-11 -0.325 x9 STORIES 0.0512 1.497 x10 VIEW 0.0556 4.39 x11 STRUCAGE -0.0037 -2.56 x12 STRUCAG2 0.00002 1.57 xx4 NEGHAGE -0.0041 -2.931 xx6 VANDAL 0.0042 2.49 xx8 BURGLAR -0.0304 -4.117 xx10 MDFAMINC -0.0000009 -0.3963 xx11 NONWHTE -0.0027 -2.766 xx12 PCTBADEG -0.0161 7.064 v1 LINKDIST v5 UNKNDUM -0.0107 -0.477 v6 NOTRADUM 0.0272 1.1.6 v7 WIN85 v8 WIN350 v9 WIN550 Constant 10.579 Model 2: Adj. R"2 = .597 Coefficients t-score 0.0438 1.702 0.0781 5.495 0.0387 2.72 0.0006 8.541 -0.0000001 -4.398 0.000001 1.047 -4E-11 -0.144 0.0515 1.506 0.057 4.537 -0.0036 -2.512 0.00002 1.547 -0.004 -2.831 0.0041 2.459 -0.0297 -4.117 -0.000001 -0.504 -0.0029 -3.007 0.0164 7.186 0.00008 3.083 10.535 Model 3: Adj. R"2 = .597 Model 4: Adj. R"2 = .597 Coefficients t-score Coefficients t-score 0.0422 1.635 0.0422 1.637 0.0784 5.494 0.0786 5.47 0.0392 2.758 0.0393 2.76 0.00061 8.512 0.0006 8.491 -0.0000001 -4.386 -0.0000001 -4.78 0.000007 0.947 0.000007 0.948 -1 E-11 -0.034 -1 E-11 -0.034 0.0528 1.538 0.0533 1.55 0.0573 4.654 0.0575 4.715 -0.0036 -2.485 -0.0036 -2.47 0.00002 1.522 0.00002 1.52 -0.004 -2.819 -0.004 -2.83 0.0042 2.436 0.0042 2.465 -0.0295 -4.112 -0.03 -4.036 -0.000001 -0.497 -0.000001 -0.495 -0.003 -3.04 -0.003 -3.007 0.0165 7.214 -0.0165 7.229 0.00093 -0.042 -0.0082 -0.38 -0.0883 -2.36 -0.094 -2.39 -0.0603 -3.265 -0.066 -2.944 -0.0153 -0.8 -0.021 -0.91 10.605 10.609 ===== PAGE 47 (score 24, numeric-paren 0, table-words 6, var-words 5) ===== Equation 1: Final Model Specification Log( sale price) = /(x1 x3 x4 x5 x6 x7 x8 x9 x10 x11 x12 xx4 xx6 xx8 xx10 xx11 xx12 v12 v13 v14) The results are shown in Table 6. Table 6: Model Results Variable Estimated Name Coefficient T-Score BATHRMS X1 4.07E-02 1.557 FIREPLC X3 7.60E-02 5.283 GARAGE X4 3.90E-02 2.733 SQFTBLD XS 6.00E-04 8.346 SQFTBD2 X6 -1.00E-07 -4.214 SQFTLOT X7 8.00E-06 1.08 SQFTLOT2 xa --6.00E-11 -0.232 STORIES X9 5.30E-02 1.547 VIEW X10 5.80E-02 4.781 STRUCAGE X11 -3.S0E-03 -2.466 STRUCAG2 X12 2.00E-05 1.48 NEGHAGE XX4 -4.20E-03 -3.001 VANDAL XX6 4.40E-03 2.58 BURGLAR xxa -3.26E-02 -4.56 MDFAMINC XX1C -1.00E-06 -0.5 NONWHTE XX1" -2.82E-03 -2.868 PCTBADEG XX1~ 1.63E-02 7.151 HIGH85 V12 -0.1022 -2.331 HIGH350 V13 '-4.63E-02 -2.237 HIGH550 V14 5.58E-03 0.26 CONSTANT 10.54 101 * .4 See footnote Coefficient P-Value lnterpretatlon* 0.12 4.070% 0 7.600% 0.006 3.900% 0 0.060% 0 0.000% 0.28 0.001% 0.817 0.000% 0.122 5.300% 0 5.800% 0.014 -0.360% 0.139 0.002% 0.003 -0.420% 0.01 0.440% 0 -3.260% 0.618 0.000% 0.004 -0.282% 0 1.630% 0.02 -9.802% 0.025 -4.545% 0.794 0.537% 4The model specified is the semi-log form so that the coefficients can be interpreted to be the percent change in price for each unit of change in the independent variable. Kennedy (1981) gives the equation for interpreting the coefficients on dummy variables in the semi- log form as g *:::::: ex~ c- V(c))-1, where g* is the true effect of the dummy independent variable on the dependent variable and V(c), is the variance of the estimated 38 ===== PAGE 48 (score 18, numeric-paren 0, table-words 9, var-words 0) ===== The above model was estimated using a heteroskedasticity- consistent covariance matrix. Earlier diagnostics on the model indicated non constant error variance across the sample, so the "hetcov" correction procedure (White 1993, pp. 79) was used with ordinary least squares regression to correct the error variance and improve confidence in both the coefficients and the t-scores. Table 6 gives the model statistics: Table 7: Model Statistics F R2 Adj. R2 ow SSE 107.059 .60 .5963 1.9076 112.96 Table 8 illustrates a simulation of the results; the marginal effect of each variable on a $135,000 house. coefficient. These values will be calculated for the coefficients of the dummy variables Table 4, Coefficient Interpretation column. 39 ===== PAGE 49 (score 11, numeric-paren 0, table-words 2, var-words 5) ===== Table 8: Model Simulation Variable Coefficient Impact on Name Interpretation $135,000 BATHRMS X1 4.0700% $5,494.5 FIREPLC X3 7.6000% $10,260.0 GARAGE X4 3.9000% $5,265.0 SQFTBLD XS 0.0600% $81.0 SQFTBD2 X6 0.0000% $0.0 SQFTLOT X7 0.0008% $1.1 SQFTLOT2 X8 0.0000% $0.0 STORIES X9 5.3000% $7,155.0 VIEW X10 5.8000% $7,830.0 STRUCAGE X11 -0.3600% -$486.0 STRUCAG2 X12 0.0020% $2.7 NEGHAGE XX4 -0.4200% -$567.0 VANDAL XX6 0.4400% $594.0 BURGLAR XX8 -3.2600% -$4,401.0 MDFAMINC XX1C -0.0001% -$0.1 NONWHTE XX11 -0.2820% -$380.7 PCTBADEG XX1~ 1.6300% $2,200.5 HIGH85 V12 -9.8019% -$13,232.6 HIGH350 V13 -4.5449% -$6,135.6 HIGH550 V14 0.5366% $724.4 40 ===== PAGE 50 (score 15, numeric-paren 0, table-words 4, var-words 7) ===== Part Five: Findings The results of the model are surprising in a couple of cases. The coefficient for stories is while only significantly different from zero at the .12 confidence level, is positive with a value of 5.3%. It was expected that with the size of structure controlled for that additional stories would lead to a reduction in price. Perhaps more surprising is the size of the coefficient for lot size, which shows an extremely small positive effect on price, although even this slight effect does not prove to be significantly different from zero. The impact of an additional bathroom on the sale price of a house is consistent with the work of Dubin and Sung (1990) who report about a 5% coefficient; Salvon (1996) finds the same effect to be, on average, 3% of sale price. In the current model, an additional bathroom is estimated to impact sale price positively at the margin by 4%. it is possible that another variable is picking up the effect. The number of fireplaces had the expected effect; the existence of one or more fire places in a house is often a proxy for housing quality or amenity level. The sign on building square feet met expectations, and the result of a negative coefficient on the squared value of building square feet is consistent with the theory of 41