# RECORD f2f2e8d4e0c68be7054271b21b3018d65c7e131c # TITLE: THE CALCULATION OF REGIONAL INDEX OF DWELLING PRICES IN CROATIA # YEAR: 2020.0 VENUE: Proceedings of the 14th Economics & Finance Conference, Lisbon DOI: 10.20472/efc.2020.014.013 # AUTHORS: Tanja Slišković; Tomislav Sekur # PAGES: 23 CANDIDATE TABLE PAGES (1-based): [19, 20] # ABSTRACT: : In the existing literature in the field of housing economics, great attention has been paid to modeling housing prices. Residential property prices are generally very important indicator as they may indicate developments in the local economy. Moreover, many existing research have proven that there is a strong interdependence of the movement of housing prices with economic developments. Therefore, different market participants (home owners, potential buyers, lessors, tenants etc.), as well as subjects from broader economic environment (banks, local government, researchers etc.) are interested in trends on housing market, and in particular in housing prices. For that reason, constant improvement of housing prices statistics is very important for all mentioned entities. In Croatia, the issue of housing prices has become topical in the last 15 years due to the turbulent developments in the domestic housing market. Generally, the indicator which is most available to the general public is average price which is published by real estate agencies and advertisers. Such indicator fails to reflect real prices movements since it encompasses the difference in quality of the offered/sold housing units on the market. The more reliable indicator would be one that is “cleared” form the impact that qualitative characteristics of the housing unit have on its price. The most commonly used method for constructing such price indicators is hedonic approach. Hedonic indices provide information about real change in the price between two points of time. The methodology of calculation hedonic housing price index in Croatia has advanced significantly over the last 12 years. In this research we use hedonic model approach to calculate the index of dwelling prices for Croatia, as well as for four regions (The City of Zagreb, Middle and South Adriatic, Istria and North Adriatic and the region Other). The variables which are included in the model are price of the dwelling, location, floor area, number of rooms, the construction year and time of sale. Two different indices based on asked and realized prices are calculated within this research. The comparison between calculated indices shows that there are large price differences between geographically different Croatian dwelling markets. Acknowledgment: This work has been supported by the Croatian Science Foundation under the project no. 6785. ===== PAGE 1 (score 7, numeric-paren 0, table-words 0, var-words 7) ===== 28 September 2020, 14th Economics & Finance Virtual Conference, Lisbon ISBN 978-80-7668-002-9, IISES DOI: 10.20472/EFC.2020.014.013 TAMARA SLIŠKOVIĆ Faculty of Economics and Business, University of Zagreb, Croatia TOMISLAV SEKUR Faculty of Economics and Business, University of Zagreb, Croatia THE CALCULATION OF REGIONAL INDEX OF DWELLING PRICES IN CROATIA Abstract: In the existing literature in the field of housing economics, great attention has been paid to modeling housing prices. Residential property prices are generally very important indicator as they may indicate developments in the local economy. Moreover, many existing research have proven that there is a strong interdependence of the movement of housing prices with economic developments. Therefore, different market participants (home owners, potential buyers, lessors, tenants etc.), as well as subjects from broader economic environment (banks, local government, researchers etc.) are interested in trends on housing market, and in particular in housing prices. For that reason, constant improvement of housing prices statistics is very important for all mentioned entities. In Croatia, the issue of housing prices has become topical in the last 15 years due to the turbulent developments in the domestic housing market. Generally, the indicator which is most available to the general public is average price which is published by real estate agencies and advertisers. Such indicator fails to reflect real prices movements since it encompasses the difference in quality of the offered/sold housing units on the market. The more reliable indicator would be one that is “cleared” form the impact that qualitative characteristics of the housing unit have on its price. The most commonly used method for constructing such price indicators is hedonic approach. Hedonic indices provide information about real change in the price between two points of time. The methodology of calculation hedonic housing price index in Croatia has advanced significantly over the last 12 years. In this research we use hedonic model approach to calculate the index of dwelling prices for Croatia, as well as for four regions (The City of Zagreb, Middle and South Adriatic, Istria and North Adriatic and the region Other). The variables which are included in the model are price of the dwelling, location, floor area, number of rooms, the construction year and time of sale. Two different indices based on asked and realized prices are calculated within this research. The comparison between calculated indices shows that there are large price differences between geographically different Croatian dwelling markets. Acknowledgment: This work has been supported by the Croatian Science Foundation under the project no. 6785. Keywords: Housing Price Index, Dwelling, Hedonic Estimation, Multiple Regression Analysis, Croatian Housing Market JEL Classification: R21, R31, O18 134 ===== PAGE 19 (score 39, numeric-paren 0, table-words 12, var-words 2) ===== Appendix Table 9:The estimation of the hedonic model of asked prices in Zagreb Dependent Variable: LOG(P1) Method: Least Squares Date: 08/30/20 Time: 00:15 Sample: 1 3337 Included observations: 3219 White heteroskedasticity-consistent standard errors & covariance Variable Coefficient Std. Error t-Statistic Prob. C 7.706989 0.017256 446.6207 0.0000 DCRNOMEREC -0.192300 0.017415 -11.04192 0.0000 DDUBRAVA -0.377987 0.019585 -19.29949 0.0000 DMAKSIMIR -0.096760 0.018481 -5.235652 0.0000 DMEDVESCAK 0.055555 0.023387 2.375410 0.0176 DNZISTOK -0.418401 0.017952 -23.30669 0.0000 DNZZAPAD -0.313912 0.017758 -17.67760 0.0000 DPESCENICA -0.251558 0.016708 -15.05572 0.0000 DPODSLJEME -0.146101 0.021798 -6.702402 0.0000 DTRESNJEVKAJUG -0.249923 0.016653 -15.00791 0.0000 DTRESNJEVKASJEVER -0.252768 0.016343 -15.46675 0.0000 DTRNJE -0.153101 0.019373 -7.902877 0.0000 DZAPAD -0.368335 0.015318 -24.04536 0.0000 DOSTALO -0.649177 0.022750 -28.53489 0.0000 M2 -2.91E-07 3.70E-07 -0.785351 0.4323 ROOMS 0.007505 0.003754 1.999280 0.0457 NEWBUILT 0.099279 0.007989 12.42750 0.0000 Y2011 -0.043895 0.009690 -4.529957 0.0000 Y2012 -0.061058 0.013966 -4.371788 0.0000 Y2013 -0.158550 0.011234 -14.11381 0.0000 Y2014 -0.205981 0.013525 -15.22945 0.0000 Y2015 -0.220215 0.018519 -11.89116 0.0000 Y2016 -0.266672 0.017488 -15.24890 0.0000 Y2017 -0.191486 0.025517 -7.504326 0.0000 Y2018 -0.106877 0.032034 -3.336331 0.0009 Y2019 -0.006022 0.043527 -0.138343 0.8900 R-squared 0.436415 Mean dependent var 7.449059 Adjusted R-squared 0.432002 S.D. dependent var 0.261592 S.E. of regression 0.197151 Akaike info criterion -0.401652 Sum squared resid 124.1069 Schwarz criterion -0.352569 Log likelihood 672.4583 Hannan-Quinn criter. -0.384060 F-statistic 98.90049 Durbin-Watson stat 1.656141 Prob(F-statistic) 0.000000 Table 10:The estimation of the hedonic model of realized prices in Zagreb Dependent Variable: LOG(P2) Method: Least Squares Date: 08/27/20 Time: 15:10 Sample: 1 3337 Included observations: 2165 White heteroskedasticity-consistent standard errors & covariance Variable Coefficient Std. Error t-Statistic Prob. C 7.623342 0.020539 371.1680 0.0000 DCRNOMEREC -0.195300 0.022228 -8.786229 0.0000 DDUBRAVA -0.361379 0.022825 -15.83260 0.0000 DMAKSIMIR -0.120617 0.023352 -5.165284 0.0000 28 September 2020, 14th Economics & Finance Virtual Conference, Lisbon ISBN 978-80-7668-002-9, IISES 152 ===== PAGE 20 (score 37, numeric-paren 0, table-words 9, var-words 2) ===== DMEDVESCAK 0.003451 0.028025 0.123140 0.9020 DNZISTOK -0.396463 0.021462 -18.47240 0.0000 DNZZAPAD -0.372616 0.023762 -15.68091 0.0000 DPESCENICA -0.236935 0.022053 -10.74384 0.0000 DPODSLJEME -0.177905 0.024790 -7.176523 0.0000 DTRESNJEVKAJUG -0.213640 0.019579 -10.91145 0.0000 DTRESNJEVKASJEVER -0.242985 0.019835 -12.25032 0.0000 DTRNJE -0.146050 0.023029 -6.342120 0.0000 DZAPAD -0.355352 0.019206 -18.50261 0.0000 DOSTALO -0.630402 0.039632 -15.90634 0.0000 M2 -8.16E-07 3.10E-07 -2.628739 0.0086 ROOMS 0.001888 0.004493 0.420138 0.6744 NEWBUILT 0.105327 0.010633 9.905290 0.0000 Y2011 -0.053690 0.010777 -4.981859 0.0000 Y2012 -0.058429 0.016597 -3.520490 0.0004 Y2013 -0.192586 0.019340 -9.958040 0.0000 Y2014 -0.198033 0.013706 -14.44818 0.0000 Y2015 -0.195098 0.018972 -10.28352 0.0000 Y2016 -0.217193 0.018910 -11.48542 0.0000 Y2017 -0.118972 0.025440 -4.676569 0.0000 Y2018 -0.050125 0.033336 -1.503644 0.1328 Y2019 0.045607 0.042373 1.076311 0.2819 R-squared 0.385131 Mean dependent var 7.347311 Adjusted R-squared 0.377944 S.D. dependent var 0.256927 S.E. of regression 0.202640 Akaike info criterion -0.342835 Sum squared resid 87.83365 Schwarz criterion -0.274621 Log likelihood 397.1193 Hannan-Quinn criter. -0.317889 F-statistic 53.59151 Durbin-Watson stat 1.802330 Prob(F-statistic) 0.000000 Table 11: The estimation of the hedonic model of asked prices in Middle and South Adriatic Dependent Variable: LOG(P1) Method: Least Squares Date: 08/30/20 Time: 01:16 Sample: 1 764 Included observations: 760 White heteroskedasticity-consistent standard errors & covariance Variable Coefficient Std. Error t-Statistic Prob. C 7.727209 0.041317 187.0236 0.0000 DUBNER 0.128713 0.030430 4.229751 0.0000 SIBKNIN -0.299180 0.033002 -9.065405 0.0000 ZAD -0.218092 0.028281 -7.711500 0.0000 M2 -0.001484 0.000729 -2.034951 0.0422 ROOMS 0.026734 0.017047 1.568189 0.1173 NEWBUILT 0.097784 0.026151 3.739135 0.0002 Y2011 -0.059602 0.037582 -1.585942 0.1132 Y2012 -0.105214 0.035846 -2.935163 0.0034 Y2013 -0.215053 0.037513 -5.732712 0.0000 Y2014 -0.175157 0.038453 -4.555087 0.0000 Y2015 -0.255112 0.067164 -3.798320 0.0002 Y2016 -0.078141 0.037563 -2.080287 0.0378 Y2017 -0.094542 0.047668 -1.983357 0.0477 Y2018 -0.007452 0.095247 -0.078238 0.9377 Y2019 -0.396220 0.162145 -2.443609 0.0148 R-squared 0.262737 Mean dependent var 7.521154 Adjusted R-squared 0.247873 S.D. dependent var 0.311264 S.E. of regression 0.269945 Akaike info criterion 0.239629 Sum squared resid 54.21538 Schwarz criterion 0.337172 28 September 2020, 14th Economics & Finance Virtual Conference, Lisbon ISBN 978-80-7668-002-9, IISES 153 ===== PAGE 21 (score 38, numeric-paren 0, table-words 12, var-words 1) ===== Log likelihood -75.05883 Hannan-Quinn criter. 0.277191 F-statistic 17.67585 Durbin-Watson stat 1.839914 Prob(F-statistic) 0.000000 Table 12: The estimation of the hedonic model of realized prices in Middle and South Adriatic Dependent Variable: LOG(P2) Method: Least Squares Date: 08/30/20 Time: 01:21 Sample: 1 764 Included observations: 760 Variable Coefficient Std. Error t-Statistic Prob. C 7.762938 0.071034 109.2849 0.0000 DUBNER 0.120050 0.042691 2.812087 0.0051 SIBKNIN -0.337109 0.059068 -5.707172 0.0000 ZAD -0.315614 0.050880 -6.203091 0.0000 M2 -0.001967 0.000884 -2.225607 0.0263 ROOMS 0.010474 0.020836 0.502682 0.6153 NEWBUILT 0.074622 0.049487 1.507921 0.1320 Y2011 -0.030996 0.071972 -0.430670 0.6668 Y2012 -0.178715 0.062808 -2.845430 0.0046 Y2013 -0.258386 0.063389 -4.076213 0.0001 Y2014 -0.221898 0.066987 -3.312550 0.0010 Y2015 -0.262983 0.092377 -2.846845 0.0045 Y2016 -0.077674 0.074019 -1.049381 0.2943 Y2017 -0.083912 0.079037 -1.061678 0.2887 Y2018 -0.009276 0.150266 -0.061731 0.9508 Y2019 -0.379294 0.258008 -1.470088 0.1420 R-squared 0.146218 Mean dependent var 7.433601 Adjusted R-squared 0.129004 S.D. dependent var 0.466968 S.E. of regression 0.435808 Akaike info criterion 1.197599 Sum squared resid 141.3070 Schwarz criterion 1.295142 Log likelihood -439.0875 Hannan-Quinn criter. 1.235161 F-statistic 8.494439 Durbin-Watson stat 1.907842 Prob(F-statistic) 0.000000 Table 13:The estimation of the hedonic model of asked prices in Istria and North Adriatic Dependent Variable: LOG(P1) Method: Least Squares Date: 08/30/20 Time: 01:34 Sample: 1 371 Included observations: 371 White heteroskedasticity-consistent standard errors & covariance Variable Coefficient Std. Error t-Statistic Prob. C 7.638522 0.058302 131.0168 0.0000 LICSENJ -0.486551 0.069408 -7.010057 0.0000 PRIMGOR 0.023283 0.071772 0.324407 0.7458 M2 -0.000919 0.000506 -1.815385 0.0703 ROOMS -0.034939 0.023296 -1.499773 0.1346 NEWBUILT 0.175797 0.042835 4.104006 0.0001 Y2011 0.010525 0.071917 0.146351 0.8837 Y2012 -0.242294 0.084286 -2.874677 0.0043 Y2013 -0.309512 0.064353 -4.809625 0.0000 Y2014 -0.161151 0.047011 -3.427920 0.0007 Y2015 -0.245795 0.070573 -3.482869 0.0006 Y2016 -0.376120 0.059516 -6.319657 0.0000 Y2017 -0.089088 0.093580 -0.952002 0.3417 28 September 2020, 14th Economics & Finance Virtual Conference, Lisbon ISBN 978-80-7668-002-9, IISES 154