# RECORD d37971d97af2ef251a0da6b9c73ae5f307074e7d # TITLE: Hedonic Analysis of Housing Prices Near the Portland Urban Growth Boundary, 1978-1990 # YEAR: 1996.0 VENUE: nan DOI: 10.15760/etd.1344 # AUTHORS: A. Alkadi # PAGES: 127 CANDIDATE TABLE PAGES (1-based): [90, 91, 97, 104, 105] # ABSTRACT: An abstract of the dissertation of Abdullah Alkadi for the Doctor of Philosophy in Urban Studies presented May 6, 1996. Title: Hedonic Analysis of Housing Prices near the Portland Urban Growth Boundary, 1978-1990. The cornerstones of Oregon's 1973 Senate Bill 100 are the preservation of farm, forest, and other resource lands and the containment of urban development within urban growth boundaries (UGB). The UGB is a boundary around each incorporated city containing enough land to meet projected needs until the year 2000. The Land Conservation and Development Commission (LCDC), charged with adopting and implementing state planning policy, sought to keep UGBs small enough to contain urban sprawl. To avoid the potential effects of land price inflation, LCOC allowed UGBs to include more land supply than the forecasted demand. The Portland-Metropolitan region was allowed to have a 15.3-percent surplus. Policy makers are unsure what effect UGBs have on housing costs. The common belief is that by restricting the amount of land available for residential construction the market drives prices up. Contrasting opinions suggest that by substituting low-density with high-density development, per-unit construction costs are lower, thus reducing the costs of owning a hom.e_. -------··. Efforts to dispel some of the mystery about the relationship between UGBs and housing prices are needed. The objective of this research is to provide empirical evidence of the relationship between the Portland-Metropolitan area's UGB and housing prices. The study uses a hedonic model to conduct a time-series analysis for the years 1978 to 1990 for Washington County. This study found no relationship between housing price and the imposition of the UGB. In fact, the rate of increase in price for single-family housing after UGB implementation was found to be much less than before. Proximity as measured by distance of sale to the UGB was the only variable that was associated with a higher rate of increase in housing prices. All of these results, with the exception of those related to proximity, were unexpected but may be explained by several factors: imposition of the Metropolitan Housing Rule in 1981, a severe recession during the 1980s, and excess land supply. These influences do not support a conclusion that UGBs lead to an increase in housing prices, at least prior to 1990, when the UGB did not constrain the ===== PAGE 1 (score 3, numeric-paren 0, table-words 0, var-words 3) ===== Portland State University Portland State University PDXScholar PDXScholar Dissertations and Theses Dissertations and Theses 1996 Hedonic Analysis of Housing Prices Near the Hedonic Analysis of Housing Prices Near the Portland Urban Growth Boundary, 1978-1990 Portland Urban Growth Boundary, 1978-1990 Abdullah Alkadi Portland State University Follow this and additional works at: https://pdxscholar.library.pdx.edu/open_access_etds Let us know how access to this document benefits you. Recommended Citation Recommended Citation Alkadi, Abdullah, "Hedonic Analysis of Housing Prices Near the Portland Urban Growth Boundary, 1978-1990" (1996). Dissertations and Theses. Paper 1345. https://doi.org/10.15760/etd.1344 This Dissertation 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 90 (score 34, numeric-paren 17, table-words 5, var-words 6) ===== 75 implementation of the UGB), and the UGB variable was eliminated to allow for perfect comparison. Table II shows the regression results. TABLE I MULTIPLE REGRESSION FOR PRICE AS A DEPENDENT VARIABLE (SAMPLE PERIOD BEFORE AND AFTER THE UGB) Variable Coefficient T-Score UGB -20583 -8.4*** AGE -290 -7.2*** BEDROOM 1781 2.1** BATHROOM 7836 7.8*** GARAGE 3312 1.9* FIREPLACE 551 0.6 INTRATE 3653 8.1 *** DCBD -0.23 -4.5*** DSCH 1.39 3.7**'" DUGB -0.68 -4.8*** DREC 0.57 1.6* DHWYS -0.28 -2.9*** DEMP -0.11 -1.2* CRIME -8370 -4.9*** EDUCATION 1.5 1.4* TIME -18.1 -0.6 TIME*LOTSIZE 0.0015 3.5*** TIME"SQFT 0.207 22.7*** Constant 34866 3.9*** Multiple R2 0.53 Sample Size 2269 .. Sigmficant at 0.05 level (one-tmled) Significant at 0.05 level (two-tailed) *** Significant at 0.01 level (two-tailed) * ** The regression results show that all the independent variables jointly explained 52, 62, and 53 percent of the variation in sale prices of homes sold before and after, on!y before, Hand only after the implementation of the UGB, respectively. ===== PAGE 91 (score 56, numeric-paren 40, table-words 4, var-words 6) ===== TABLE II MULTIPLE REGRESSION FOR PRICE AS A DEPENDENT VARIABLE (SAMPLE PERIOD BEFORE AND AFTER THE UGB) Variable Before-After UGB Before UGB Coeff. T-Score Coeff. AGE -287 -7.0*** -140 BEDROOM 1754 2.0** 149 BATHROOM 7695 7.6*** 11589 GARAGE 3001 1.7* 1741 FIREPLACE 502 0.5 -1628 INTRATE 1884 4.6*** -430 DCBD -0.23 -4.4*** -0.11 DSCH 1.3 3.5*** 0.6 DUGB -0.67 -4.7*** -0.50 DREC 0.61 1.7* 0.39 DHWYS -0.26 -2.7*** -0.43 DEMP -0.09 -1.0* 0.25 CRIME -6560 -3.8*** -1240 EDUCATION 1.5 1.4* 1.8 TIME -188 -8.7*** -461 TIME*LOTSIZE 0.0015 3.6*** 0.013 TIME*SQFT 0.207 22.4*** 0.547 Constant 48478 5.5*** 51125 Multiple R2 0.52 0.62 Sample Size 2269 335 .. Significant at 0.05 level (one-tailed) Significant at 0.05 level (two-tailed) *** Significant at 0.01 level (two-tailed) ** T-Score -1.9* 0.1 6.9*** 0.6 -0.9 -0.3 -1.4* 0.3 -2.1** 0.6 -2.9*** 1.6 -0.5 1.1* -1.4 6.0*** 6.6*** 3.3*** After UGB Coeff. T-Score -341 -7.6*** 1222 1.2* 6277 5.5*** 3820 1.9* 726 0.7 4454 5.7*** -0.23 -4.0*** 1.3 3.3*** -0.65 -4.1*** 0.64 1.6* -0.27 -2.5*** -0.15 -1.5* -10326 -3.4*** 0.87 0.6 11 0.3 0.0013 2.9*** 0.217 21.9*** 11892 0.7 0.53 1934 76 Comparing the three regressions, it is obvious that some variables became significant after the implementation of the UGB while others did not. Although the coefficients of DUGB are significant before and after the implementation of the UGB at a level of .05 and .01, respectively, the coefficient after the implementation of the UGB is higher than before the UGB was implemented, a finding which is consistent with the theoretical literature (Correll et al. 1978; Nelson 1984). In particular, the coefficients of DUGB suggest that a house close to the UGB after it was implemented will lose $0.65 for ===== PAGE 92 (score 8, numeric-paren 0, table-words 2, var-words 4) ===== 77 every foot of distance from the UGB while the value before the implementation was $0.50 for every foot of distance. It must be kept in mind, however, that the period after the UGB was implemented experienced a severe depression. Thus, the $0.65 could be underestimated. When looking to the rate of increase it is clear that a house sold after the UGB was implemented captured a higher rate of increase, but it is not that high compared to the rate of increase before the UGB was implemented. In particular, due to DUGB, a house after the UGB was implemented captured 0.00080 percent of the house value for every single foot it gets closer to the UGB versus 0.00071 before the UGB was implemented. Nevertheless, it could also be argued that the DUGB before the implementation of the UGB was significantly sizable, due to the fact that participants in the land market feared a stringent UGB and its consequences and were unsure of the final lines of the UGB, drafted before October 1980. In fact, the date of line demarcation goes back to 1977 and that is how Knaap (1982) found some effects of the UGB on land values when he studied the Portland metropolitan area. Based on the coefficients of the two interactions, TIME*LOTSIZE and TIME*SQFT, as shown in Table II, the real implicit price trends in house value can be calculated for each period. The coefficients of the interactions show the price increment of one square foot in LOTSIZE and SQFT. Multiplying these coefficients with the average size of LOTSIZE and SQFT in a certain period would reveal the real implicit price of an average LOTSIZE per month, as the TIME variable was measured in a one-month increment. In particular, the average LOTSIZE _Q_tJring_the __ J>erio_d after UGB was 11,827 square feet. Multiplying the coefficient outcome for the same variable during the same period reveals $15. This is the increment in ===== PAGE 97 (score 49, numeric-paren 34, table-words 3, var-words 6) ===== 82 compared to the preceding and following periqds. In fatt, this is not unexpected, as the results are consistent with the depression that occurred in the Portland ~lousing market duri1;1g the period between 1983-1985. TABLE IV MULTIPLE REGRESSION FOR PRICE AS A DEPENDENT VARIABLE (SAMPLE PERIOD AFTER THE UGB) I Variable 1980-1982 1983-1985 Coef. T-Score Coef. T-Score AGE -442 -2.1** -160 -1.6* BEDROOM -3308 -0.8 7194 3.1*** BATHROOM 15800 3.1*** 4280 1.7* GARAGE -10106 -1.2 3077 0.7 FIREPLACE 3107 0.7 2058 0.8 INTRATE -2155 -0.6 2459 0.6 DCBD 0.09 0.3 -0.17 -1.4* DSCH 2.4 1.4* 0.6 0.7 DUGB 0.1 0.2 -0.35 -0.9 DREC -0.9 -0.6 O.Dl 0.01 DHWYS -0.09 -0.2 -0.07 -0.2 DEMP 0.003 0.00 0.004 0.01 CRIME -25861 -0.8 -103321 -2.0** EDUCATION -0.1 -0.03 -0.9 -0.3 TIME -945 -1.2 67.4 0.2 TIME*LOTSIZE 0.009 0.8 -0.002 -2.4*** TIME*SQFT 0.465 4.7••• 0.224 7.1*** Constant 146284 1.4* 189726 2.3** Multiple R2 0.50 0.41 Sample Size 178 400 .. Significant at 0.05 level (one-tailed) Significant at 0.05 level (two-tailed) ••• Significant at 0.01 level (two-tailed) * •• 1986-1987 Coef. T-Score -29$ -2.9*** 689() 3.2*** 646$ 2.7*** 353,5 0.7 -2194 -1.0 6537 0.9 -0.3 -2.4 .... 0.6 0.7 -0.8~~ -2.8*** 0.54 0.6 -0.28 -1.3* -0.1 -0.6 -1947 -0.04 5.3 1.8* 371 0.8 0.00~ 2.4*** 0.16;~ 7.6*** -7341)4 -0.5 0.47 459 1988-1990 Coef. T-Score -488 -8.5*** -5261 -3.8*** 4855 3.1*** 5586 2.1** 988 0.7 -10566 -4.0*** -0.26 -3.2*** 1.2 2.1** -0.68 -2.0*** 1.3 2.4*** -0.37 -2.5*** 0.2 1.4 -1920 -0.2 ' 0.5 0.2 I 768 3.6*** I 0.002 4.2*** 0.252 21.5*** 45390 1.1* I 0.67 i 897 However, the four regressions show that all the independent variables jointly explained 50, 41, 47, and 67 percent of all the variations in sale prices of homes sold after UGB implementation and ===== PAGE 98 (score 9, numeric-paren 0, table-words 3, var-words 2) ===== 83 between the periods 1980-1982, 1983-1985, 1986-1987, and 1988-1990, respectively. The interactive eff~cts of LOTSIZE*TIME and SQFT*TIME are consistent with previous theoretical predictions. All of their coefficients are significant at least at the .01 level except for TIME*LOTSIZE for the period between 1980-1982, which is not significant. Based on earlier discussion of the coefficients of the two interactions TIME*LOTSIZE and TIME*SQFT, the real implicit price trends of the price of a house when looking at the components of LOTSIZE and SQFT are summarized in Table V and Figure 14. LOTSIZE and SQFT fluctuated and SQFT decreased during the two periods followed the first one and then picked up. LOTSIZE followed a zigzag pattern decreasing and increasing and then decreasing again. TABLE V REAL IMPLICIT PRICE TRENDS (IN $) PER MONTH DUE TO LOT SIZE AND HOUSE IMPROVEMENTS FOR THE PERIODS AFTER THE UGB 1980-1982 1983-1985 1986-1987 1988-1990 Value % Value % Value o/o Value 0/o TOTAL 865 1.09 361 0.49 321 0.42 482 0.53 LOTSIZE 74 0.09 -32 -0.04 33 0.04 26 0.03 SQFT 791 1.0 393 0.53 288 0.38 456 0.50 In particular, during the period between 1983-1985, both interactions TIME*LOTSIZE and TIME*SQFT obtained lower coefficients compared to the preceding period. Again, this was expected due to the depression thafoccurredduring this period. What is interesting is that during the period of 1986 and 1987 TIME*LOTSIZE picked up in its coefficient while TIME*SQFT became ===== PAGE 104 (score 56, numeric-paren 42, table-words 3, var-words 6) ===== TABLE VI MULTIPLE REGRESSION FOR PRICE AS A DEPENDENT VARIABLE (SAMPLE PERIOD BEFORE AND AFTER THE UGB) Variable Before-After UGB Before UGB Coeff. T-Score Coeff. AGE -323 -7.7*** -173 BEDROOM 1849 2.1** 620 BATHROOM 7587 7.5*** 10512 GARAGE 2630 1.4* 2280 FIREPLACE 513 0.5 -296 INTRATE 1915 4.7*** -1345 DCBD -0.22 -4.2*** -0.13 DSCH 1.2 3.3*** 1.3 DUGB -0.67 -4.7*** -0.62 DREC 0.48 1.3* 0.33 DHWYS -0.26 -2.7*** -0.43 DEMP -0.1 -1.1 * 0.17 CRIME -6464 -3.7*** -2841 EDUCATION 1.7 ! 1.5* 1.5 TIME -173 -8.0*** -203 TIME* LOT ANfT 75 I 5.3*** 506 TIME*SQFT 0.204 22.1*** 0.587 Constant 48302 5.5*** 61092 Multiple R2 0.52 0.60 Sarnele Size 2269 335 . ' .. S1gmficant at 0.05 level (one-tailed) Significant at 0.05 level (two-tailed) *** Significant at 0.01 level (two-tailed) ** T-Score -2.2** 0.4 6.1 *** 0.8 -0.1 -0.9 -1.5* 1.9* -2.5** 0.5 -2.8*** 1.2* -1.2* 0.9 -0.6 4.3*** 6.9*** 3.8*** After UGB Coeff. T-Score -390 -8.5*** 1335 1.3* 6229 5.5*** 3320 1.6* 686 0.7 4644 6.0*** -0.22 -3.8*** 1.2 3.0*** -0.63 -4.0*** 0.48 1.2* -0.27 -2.5*** -0.16 -1.5* -9765 -3.2*** 1.1 0.8 34 0.9 77 5.3*** 0.211 21.5*** 8014 0.4 0.53 1934 89 The coefficients for TIME*LOTAMT, the interaction variable, are significant during the three periods following 1982, and, in fact, for these three periods, the level of significance went up as time proceeded, to the .05 level (one-tail), .05 level (two-tail), and 0.01 level (two-tail), re~pectively. i The magnitudes of the coefficients also increased as Hme pasc::ed A honse ~~ a---largeHlot-during 1983-1985 captured $58 per month, $66 per month during 1986-1987, and $75 per month during 1988-1990; This implies that larger lots got more scarce ===== PAGE 105 (score 49, numeric-paren 34, table-words 3, var-words 6) ===== 90 as time passed especially since the UGB prevented any inclusion of raw land since October 1980. Further details show that a house located on a large lot captured att increase of $6,372 between 1983 and 1990. However, when looking at the rate of the increase in LOTAMT price, the analysis shows that the rate was increasing until the last period, where the rate dropped. In particular, the rate was 0.025, 0.079, 0.088, and 0.083 percent, respectively. TABLE VII MULTIPLE REGRESSION FOR PRICE AS A DEPENDENT VARIABLE (SAMPLE PERIOD AFTER THE UGB WAS IMPLEMENTED) Variable 1980-1982 1983-1985 1986-1987 Coef. T-Score Coef. T-Score Coef. T-Score AGE -390 -1.8* -216 -2.1" -293 -2.8*** BEDROOM -3127 -0.8 8005 3.5*** 7115 3.3*** BATHROOM 15922 3.1 *** 4514 1.8* 6743 2.8*** GARAGE -9105 -1.1 2150 0.5 5137 1.1* FIREPLACE 3358 0.8 1226 0.5 -2054 -0.9 INTRATE -2334 -0.7 1925 0.4 6816 0.9 DCBD 0.08 0.3 -0.21 -1.6* -0.3 -2.3** DSCH 2.7 1.6* 0.03 O.Q3 0.6 0.7 DUGB 0.1 0.1 -0.35 -0.9 -0.8 -2.7*** DREC -0.9 -0.6 0.14 0.16 0.59 0.7 DHWYS -0.1 -0.2 -0.07 -0.3 -0.27 -1.2* DEMP -0.001 -0.00 0.05 0.2 -0.1 -0.6 CRIME -23783 -0.7 -104495 -2.0** -1269 -0.02 EDUCATION -0.4 -0.09 -0.9 -0.3 5.1 1.7* TIME -822 -1.1 81 0.2 417 0.9 TIME*LOT AMT 20 0.1 58 1.3* 66 2.2** TIME*SQFT 0.481 4.8*** 0.179 6.0*** 0.159 7.4*** Constant 139224 1.4* 203406 2.4*** -83479 -0.5 Multiple R2 0.49 0.41 0.47 Sample Size 178 400 459 * .. ** Sigmficant at 0.05 level (one-tmled) Significant at 0.05 level (two-tailed) *** Significant at o~oriev·er (two=tailed) _________ . 1988-1990 Coef. T-Score -529 -8.9*** -4814 -3.5*** 4276 2.8"** 4801 1.8* 835 0.6 -10200 -3.9*** -0.24 -3.1 **" 1.1 2.0** -0.72 -3.2*** 1.1 2.0** -0.39 -2.7*** 0.2 1.5 -130 -0.01 0.5 0.3 811 3.8*** 75 4.5*** 0.250 21.4*** 35436 0.9 0.67 897 ===== PAGE 106 (score 3, numeric-paren 0, table-words 0, var-words 3) ===== 91 SUMMARY By exploring the relationship between housing prices and the UGB, it was found that the UGB did not affect housing prices. In particular, the increased rate in the value of the LOTSIZE was much lower during the period after the UGB was implemented compared to the period before the UGB. In addition, during the period after the UGB was implemented, the rate of increase in lot size is almost flat. This low rate could be attributed to the ample amount of land supplied within the UGB as this analysis studied only up to the midpoint of the designated period. The analysis showed an overall increase in the rates of the value of LOTAMT during the three periods following the UGB implementation, while the rate decreased at the fourth period. Large lots did capture higher values as time passed compared to small lots, but the rate of increase during the period after the UGB was implemented was much lower compared to the period before the UGB was implemented. Finally, it has been demonstrated that proximity to the UGB was seen as an amenity and that housing prices do decline as they get away farther from the UGB. ------· -· ··--------··. ---·---------------