1# RECORD 8b40d59c2cf1fb45e2898b8a9d2bc70967606aa72# TITLE: Indirect Construction of Hedonic Price Indexes for Private Properties3# YEAR: 2008.0 VENUE: nan DOI: 10.1007/bf033992684# AUTHORS: S. Fahrländer5# PAGES: 24 CANDIDATE TABLE PAGES (1-based): [20, 21]6# ABSTRACT: SummaryThis paper analyzes the construction of price indexes for condominiums and single family houses in Switzerland over the period 1985 to 2004 using the indirect method. We find, that the pooling of the data, i.e. direct indexes with fixed hedonic prices, are good models for average properties but can be biased for the indexation of non-average properties. Therefore indirect index construction using predictions for specific properties based on annual equations is recommended.Based on differentiation of forty regions it is shown, that indexes vary from region to region, but show a comparable general path between 1985 and 2004. In recent years prices for single family houses raised in the large urban areas and tourist resorts but, due to big land reserves, were stable or even declined in smaller urban and semi-urban as well as in rural areas. Price paths for condominiums are generally comparable to those of single family houses with the big difference, that over the last three years prices were raising in most of the regions.789===== PAGE 1 (score 6, numeric-paren 0, table-words 0, var-words 6) =====1011© Swiss Journal of Economics and Statistics 122008, Vol. 144 (4) 607–63013a 14Fahrländer Partner AG, Uetlibergstrasse 20, 8045 Zürich, Switzerland, +41 43 333 05 55; sf@15fpre.ch.16 17This paper benefited from helpful comments by Robert Leu, Dominik Matter, Klaus Neus ser 18and the editors. The usual disclaimer applies. Special thanks to the data providing Swiss banks 19and insurance companies.201 21See for example Bender, Gacem and Hoesli (1994), Maurer, Pitzer and Sebastian (2000) 22or Wolverton and Senteza (2000).232 24As of the 2000 census the home ownership rate in Switzerland was at 34.6%. Today’s rate 25might be 38% to 39% and possibly increase to about 40% by 2010 (see BAK & fpre 2007, 26p. 12).27Indirect Construction of Hedonic Price Indexes 28for Private Properties29Stefan S. Fahrländera30JEL-Classification: C31, R3131Keywords: Hedonic price, private property, direct index, indirect index321. Introduction33Hedonic models have been widely used in Switzerland by mortgage lenders in 34day to day business for more than a decade. Although estimating today’s market 35values, the companies compile an increasing number of “historic” transaction 36prices together with the characteristics of the private properties. These data allow 37analysis on the construction and estimation of price path of private properties 38in Switzerland.39For modelling the price paths several methods are used in the literature. Pop-40ular but little quality adjusted and therefore susceptible to biases is the indexa-41tion of mean or median prices of sold properties. Recommended are repeat sales 42methods, since the changes of the prices of identical properties are compared.1 43However the problems of ageing and the declining condition or the considera-44tion of refurbishments remain a challenge. In addition, in markets like Switzer-45land with a low home ownership rate and a very small turnover rate of the prop-46erties the use of the repeat sales method is hardly possible since data is simply 47not available.2 Therefore the hedonic method is very popular and widely used in 484950===== PAGE 20 (score 32, numeric-paren 16, table-words 6, var-words 1) =====5152626 53Stefan S. Fahrländer54Appendix55Table A: Robust Estimation Output for Condominiums 2004a56Coefficient57Standard error58Sign. level59Intercept60373.0556172.76462***63macro640.948650.00866***67DNWF5068−1.272690.15270***71LNWF50721.114730.03574***75DNWF507576−0.352770.16478**79LNWF5075800.881810.03482***83DNWF750084−0.056850.17986n.s.87LNWF7500880.820890.03690***91LNWF0150920.810930.01794***95DNWF150P96−0.442970.14698***99LNWF150P1000.8991010.024102***103bauj104−0.6521050.073106***107bauj21080.0001090.000110***111stand: 41120.132 1130.003114***115stand: 51160.2761170.006118***119micro: 41200.1191210.003122***123micro: 51240.2831250.005126***127S.E. residuals1280.120129D.F. residuals13010 345131R2 b1320.965133Note: “.” represents a significance level of 10%, * a 5%, ** a 1% and *** a 1‰ significance level. 134DNWF50 is the dummy for condominiums with a dwelling up to 50m2, LNWF50 the respec-135tive slope. Also see table 1.136a 137For practical reasons, the quarterly coefficients for the regions are not shown.138b 139The robust R2 is calculated as proposed by Rousseeuw and Leroy (1987).140141142===== PAGE 21 (score 30, numeric-paren 13, table-words 6, var-words 1) =====143144Indirect Construction of Hedonic Price Indexes for Private Properties 145627146Table B: Robust Estimation Output for Single Family Houses 2004a147Coefficient148Standard error149Sign. level150Intercept15110.19415286.720153n.s.154macro1550.8921560.010157***158DVOL600159−1.5661600.144161***162LVOL6001630.7471640.023165***166LVOL13001670.5071680.004169***170DVOL2300171−0.7601720.373173*174LVOL23001750.6271760.051177***178LLA20001790.0021800.002181n.s.182LLAE2501830.0211840.001185***186DLAE600187−0.3011880.085189***190LLAE6001910.0771920.014193***194DLAE18001950.0551960.126197n.s.198LLAE18001990.0232000.019201n.s.202bauj203−0.1462040.087205.206bauj22070.0002080.000209.210stand: 42110.1232120.004213***214stand: 52150.2152160.007217***218micro: 42190.1092200.004221***222micro: 52230.2892240.006225***226S.E. residuals2270.121228D.F. residuals2297 163230R2 b2310.937232Note: “.” represents a significance level of 10%, * a 5%, ** a 1% and *** a 1‰ significance level. 233DVOL600 is the dummy for single family houses with a cubic content up to 600m3, LVOL600 234the respective slope. Also see table 1.235a 236For practical reasons, the quarterly coefficients for the regions are not shown.237b 238The robust R2 is calculated as proposed by Rousseeuw and Leroy (1987).239240241===== PAGE 22 (score 0, numeric-paren 0, table-words 0, var-words 0) =====242243628 244Stefan S. Fahrländer245References246ARE (2005), Raumentwicklungsbericht 2005, Ed.: Eidgenössisches Departement 247für Umwelt, Verkehr, Energie und Kommunikation; Bundesamt für Raum-248entwicklung, Bern.249BAK & fpre (2007), Wohnimmobilien 2007–2022, Ed.: BAK Basel Economics 250and Fahrländer Partner, Basel, Zürich.251Bender, A. R., B. Gacem and M. Hoesli (1994), «Construction d’indices 252immobiliers selon l’approche hédoniste», in Finanzmarkt und Portfolioma-253nagement, vol. 8, no. 4, pp. 522–533.254Bignasca, F., et al. (1996), Immobilienmarkt Zürich: Immobilienpreise und Bau-255investitionen unter der Lupe, Ed.: Zürcher Kantonalbank, Zürich.256Boskin, Michael J., et al. (1996), (Advisory Commission To Study The Con-257sumer Price Index); Toward a More Accurate Measure of the Cost of Living: 258Final Report to the Senate Finance Commission, Washington.259Brachinger, H. W., B. Schips and W. Stier (1999), „Revision Landesindex 2602000: Expertise zur Relevanz des Boskin-Reports für den schweizerischen 261Landesindex der Konsumentenpreise“, in BFS aktuell, Neuchâtel.262Cox, T. F., and M. A. A. Cox (1994), Multidimensional Scaling, Monographs on 263Statistics and Applied Probability 59, London.264Fahrländer, S. (2001a), „Hedonische Handänderungspreisindizes: Konzeption 265und Methodik“, Ed. Wüest & Partner, Zürich (mimeo).266Fahrländer, S. (2001b), „Performance von Wohneigentum“, in Immo-Moni-267toring 2002, Band I, Wohnungsmarkt, Ed.: Wüest & Partner, Zürich.268Fahrländer, S. and U. Hausmann (2001), „Ronco ist nicht Zug, Zug ist nicht 269Genf: Grosse Unterschiede in der regionalen Wertentwicklung von Wohnei-270gentum“, in Neue Zürcher Zeitung, no. 151, Zürich.271Fahrländer, S. (2006), “Semiparametric Construction of Spatial Generalized 272Hedonic Models for Private Properties”, in Swiss Journal of Economics and 273Statistics, vol. 142 (4), pp. 501–528.274Fahrländer, S. (2007), Hedonische Immobilienbewertung: Eine empirische Unter-275suchung der Schweizer Märkte für Wohneigentum 1985 bis 2005, München.276Greene, W. H. (1997), Econometric Analysis, 3rd edition, London etc.277Hampel F. R., et al. (1986), Robust Statistics: The Approach Based on Influence 278Functions, New York.279Hastie, T., and R. Tibshirani (1990), Generalized Additive Models, London.280Hastie, T., R. Tibshirani and J. Friedman (2001), The Elements of Statistical 281Learning, New York.282