1# RECORD 5bf64585206ad7e7f4dd7182eea5bcbb352319dd2# TITLE: Ripple effect on housing prices. Evidence from tourist markets in Alicante, Spain3# YEAR: 2017.0 VENUE: nan DOI: 10.3846/1648715x.2016.12411924# AUTHORS: Paloma Taltavull de La Paz; E. Lopez; F. Juárez5# PAGES: 14 CANDIDATE TABLE PAGES (1-based): [10]6# ABSTRACT: ABSTRACTThis paper analyses the house price diffusion effect in an economic-mixed region where the costal amenities strongly attracts second home and temporal residents while the main region’s city is an administrative centre in alicante province, Spain. The region is called Vega Baja country with 2 well known foreign-Europeans destination areas are located (Orihuela costa and Torrevieja). Using geo-referenced data, the paper explores the ripple effect on house prices between the coastal and inland areas, versus Orihuela capital. To control by heterogeneity and spatial autocorrelation, the model estimates housing prices controlled by quality including 33 house characteristics and spatial autocorrelation applying an SAR-hedonic based model which is estimated yearly for the period 2007–2012. Once controlled by quality, the estimated prices are used to seek 3 evidences of ripple effect: with spatial contiguity (spatial diffusion in the short distance), without special contiguity (long distance) and constant ...789===== PAGE 1 (score 8, numeric-paren 0, table-words 0, var-words 8) =====1011Copyright © 2017 Vilnius Gediminas Technical University (VGTU) Press12http://www.tandfonline.com/tspm13* Corresponding author. E-mail: paloma@ua.es14RIPPLE EFFECT ON HOUSING PRICES. EVIDENCE FROM 15TOURIST MARKETS IN ALICANTE, SPAIN16Paloma TALTAVULL de LA PAZ a,*, Enrique LÓPEZ a, Francisco JUÁREZ a17a Department of Applied Economics, University of Alicante, Campus de San Vicente del Raspeig s/n, 1803080 Alicante, Spain19Received 17 June 2015; accepted 28 October 201520Abstract. This paper analyses the house price diffusion effect in an economic-mixed region where 21the costal amenities strongly attracts second home and temporal residents while the main region’s city 22is an administrative centre in Alicante province, Spain. The region is called Vega Baja country with 232 well known foreign-Europeans destination areas are located (Orihuela costa and Torrevieja). Using 24geo-referenced data, the paper explores the ripple effect on house prices between the coastal and in-25land areas, versus Orihuela capital. To control by heterogeneity and spatial autocorrelation, the model 26estimates housing prices controlled by quality including 33 house characteristics and spatial autocor-27relation applying an SAR-hedonic based model which is estimated yearly for the period 2007–2012. 28Once controlled by quality, the estimated prices are used to seek 3 evidences of ripple effect: with spa-29tial contiguity (spatial diffusion in the short distance), without special contiguity (long distance) and 30constant relative prices ratio, using Lagrange Multiplier test and Moran’s I, and Spatially Weighted 31Two Stage Least Squares (SW2SLS) including spatial lag and errors in the model. Results show the 32existence of spatial autocorrelation patterns in a well-defined local clusters and highlight evidence of 33ripple effect between Orihuela city and the coastal and inland areas but with lagged effect.34Keywords: Housing prices; Ripple effect; Tourist areas; Spain35SUPPLEMENTARY MATERIAL associated with this article can be found, in the online version, at 36https://doi.org/10.3846/1648715X.2016.1241192371. INTRODUCTION38House prices have received increasing attention 39since the start of the Global Financial Crisis (GFC) 40because many has seen in the house price sinking 41to have part of the crisis responsibility. The EU 42authorities included house price growth rate into 43the list of indicators to be followed to monitoring 44(macro) economic imbalances and it is also recog-45nized that house prices have played a role in the 46lack of household affordability to housing there-47fore, having social impacts. This interest grew af-48ter a long list of analysis deepening on how the 49housing bubble formation happened, and giving 50different evidence about how the three different 51groups of variables (the general-global influences, 52the specific-local demand shocks and the diffusion 53effects from other closer areas) affected to prices.54It is documented how housing (aggregate) pric-55es depend on the fundamental drivers associated 56to the economic and social dynamism (for instance 57migration, economic growth or the financial sys-58tem efficiency) although a relevant part of those 59influences happen at local level as a response to 60a demand shock. Local house prices also diffuse 61and interact to other prices in the closer markets 62transmitting their influences among the territory 63in the long run. It is recognized that housing prices 64maintain a weighted relationship in the long run 65with other regions’ ones showing a long term equi-66librium in the relative house price among regions 67in the long run (Meen 1999; Holmes, Grimes 2008 68among others).69The literature agrees that house prices depend 70on the local demand characteristics and that dif-71fer across markets. A controversy opinion appears 72International Journal of Strategic Property Management73ISSN 1648-715X / eISSN 1648-9179742017 Volume 21(1): 1–1475doi:10.3846/1648715X.2016.1241192767778===== PAGE 10 (score 39, numeric-paren 0, table-words 12, var-words 5) =====79801081P. Taltavull et al.82three spatial areas, with high explanatory capacity 83(AdjR2 > 0.8). The three regions show consistent he-84donic explanatory structure, with larger relevance 85of the neighbourhood quality in Orihuela and the 86coastal area, and more relevance of population dy-87namism and the use density to explain house prices 88in the inland region. The spatial lag parameters are 89significant in all equations and consistent results 90are at region level as well as in the case of spatial 91error parameter. Those agrees with the literature 92which identify the significance of spatial lag param-93eters as evidence of ripple effect, both in prices (rho 94parameter for spatial lagged prices) and for unob-95served heterogeneity capturing spatial-other-influ-96ences (lambda parameter for spatial error, Cameron 97et al. 2005). Spatial error parameter is not signifi-98cant in two years (2009 and 2012) for Orihuela, and 99in one (2012 and 2011) for the coastal area and the 100inland region suggesting that for those periods, the 101influence from unobserved variables at spatial level 102have not effect on housing prices.103The last step seeks to find the second evidence 104of ripple effect through fitting the model (7), that 105is, estimating the parameters which allow ap-106proach the ratio to capture the constant relation-107ship among the prices. Constancy and convergence 108to the equilibrium should be an evidence of the 109existing ripple effect between the coast and Ori-110huela capital, in one hand, the inland to Orihuela 111capital, on the other hand.112Table 2 shows the estimated parameter and the 113goodness of fit of time hedonic equation. The mod-114el has been fitted between two price series, taking 115Orihuela prices as reference. A third model is also 116estimated calculating the conditional parameters of 117prices in the three spatial dimensions relative to the 118prices in Orihuela. Those parameters serve to calcu-119late the “k” values as explained above. The estimated 120parameter values are consistent across the years.121The “k” values are reported in Table 3 and also 122reports the values of the parameters’ exponents. 123Note that the joint model contains the conditional 124Table 2. Relative price models estimation. Parameters value and explanatory capacity125log pi = g + x(log pj). Dependent variable log(porihuela)126200712720081282009129201013020111312012132log porihuela = gc + xc (log pcoast)133gc1342.4151353.1621363.1981375.0371382.0321392.291140xc141–0.207142–0.273143–0.278144–0.436145–0.179146–0.201147Adj R21480.1011490.14071500.1661510.3211520.1111530.124154log porihuela = gin + xin (log pinland)155gin1562.5081572.9071582.4591592.5621601.3861611.564162xin163–0.217164–0.254165–0.217166–0.227167–0.123168–0.139169Adj R21700.1101710.1151720.0881730.0561740.0471750.052176Conditional coast and inland joint model177log porihuela = g(c+int) + x int/c (log pinland) + x c/int (log pcoast)178g(c+int)17910.2771809.8751819.65718210.3161839.3341849.947185x c/int186–0.880187–0.852188–0.841189–0.893190–0.823191–0.873192x int/c193–0.891194–0.863195–0.854196–0.913197–0.831198–0.886199Adj R22000.8822010.8292020.8462030.8852040.8292050.877206Table 3. Relative price model constancy estimation “k” value207200720820082092009210201021120112122012213Mean214‘k’ values215Coastal area [e gc * e xc]2160.91852170.57912180.56122190.13552201.13662210.97592220.717223Inland [e gint * e xint]2240.86362250.6742260.87992270.82672281.7342291.52452301.083231Conditional Coastal and inland232[e g(c+int) * e x c/int * e x int/c]2331.4E-082344.9E-082358.6E-082368.9E-092372.2E-072382.8E-082396,7E-08240Exponent of price parameter coefficients241Coastal area [e xc]242–0.499243–0.862244–0.890245–2.197246–0.364247–0.461248–0.871249Inland [e xint]250–0.545251–0.739252–0.535253–0.581254–0.171255–0.218256–0.464257Model with conditional parameters258Coastal area [e x c/int]259–9.044260–8.411261–8.122262–9.215263–7.680264–8.688265–8.526266Inland [e x int/c]267–9.159268–8.524269–8.251270–9.414271–7.757272–8.817273– 8.653274275276===== PAGE 11 (score 7, numeric-paren 1, table-words 0, var-words 5) =====27727811279Ripple effect on housing prices. Evidence from tourist markets in Alicante, Spain280parameters of prices between cost/inland to Ori-281huela which capture the multiple effect among 282prices in those areas.283Figure 4 represents the two “k” values-series 284for coastal/Orihuela and Inland/Orihuela, both by 285year and the period mean. As it can be seen, there 286is a dissimilar evolution of both relative ratios dur-287ing the analysed period. The first (coast/Orihuela) 288ratio returns to the equilibrium at the end of the 289period after to diverge since 2007, reducing from 2900.9 to 0.6, and in 2010, with large changes mak-291ing the ratio become highly volatile to return to 292the equilibrium in 2012. Those “k” evolve around a 293mean value of 0,71 and reverts to the mean during 294the period10.295In the case of second ratio (inland/Orihuela), 296the “k” is shown as a constant until 2010 when 297starts diverging until the end of period with no 298apparent convergence. Both periods match with 299the double-v crisis experience in Spain with a first 300shock, the GFC crisis starting in 2007, and the 301“austerity crisis” starting in 2010–11. The latter 302did strongly negatively impact on Spanish econo-303my and especially in the households demand. The 304first period seems to suggest equilibrium in house 305prices between both regions until 2010 while the 306second shows a distortion on the price relations 307where the “k” parameter strongly differs from the 308mean (1.08).309The model with conditional parameters gives 310an estimator of a combined “k” with a mean of 3116,7*e(-8), showing the same shocks effect on house 312relative prices explained above (increasing volatil-313ity since 2010), but clearly converges to the equi-314librium (Fig. 5) at the end of the period. Conver-315gence to the equilibrium is also achieved by the 316estimated parameters values and, as an illustra-317tion, is represented in Figure 7.318The results obtained give empirical evidence of 319convergence relationships among housing prices 320between areas under study which shows different 321socio-economic characteristics although located at 322close distance. The analysed period shows the ef-323fect of shocks on relative house prices which are 324consistent with the 2 waves of crisis affecting 325Spanish economy. The change in “k” value during 326the first years report a reduction in house prices 32710 The sample period, 2007 to 2012, is too short to con-328firm without doubt that mean “k” is a good benchmark 329of long run equilibrium. However, the empirical exer-330cise gives evidence that it occurs during the analysed 331period. Further analysis with longer database should 332be develop to support the evidence in this paper. We 333acknowledge to an anonymous referee to focus in this 334point.335in the costs rather than in the capital. This sug-336gest that the impact of GFC shock at worldwide 337level (1st shock) would have reduced income (or 338purchase capacity) of foreign demander before the 339Spanish ones, lowering the foreign house demand 340and diminishing house prices in the coast where 341the demand is dominated by foreign (European) 342income level and flow of population, and modifying 343the relative to Orihuela house prices, where house 344market is reflecting the local socio-demographic 345characteristics.346The second shock (the one coming from the 347application of austerity measures in the Spanish 348Economy) had domestic effects, that is, it affected 349the income level to Spanish households influenc-350ing house prices dominated by domestic demand to 351lower. As the fall was sharp, it is consistent with 352how the relative “k” ratio jumps up and return to 353the equilibrium (average “k”) quickly. This is con-354sistent with results in Cook (2003, 2005).355Although the period of time is short to observe 356the full equilibrium, the described evolution seems 357to suggest the existence of ripple effect between 358the coast (dominated by foreign demand) and the 3590.000 3600.200 3610.400 3620.600 3630.800 3641.000 3651.200 3661.400 3671.600 3681.800 3692.000 370200737120083722009373201037420113752012 376Coastal area377 378Interior 379Fig. 4. Constant price relationship and convergence380γ381ξ382γ383ξ384385386387=388=389390391392 393 394* 395, 396 397* 398c399c400int401int402Coastalarea403e404e405Interior406e407e4080.0004090.0504100.1004110.1504120.2004130.250414200741520084162009417201041820114192012420Fig. 5. Constant relationship and convergence. ‘k’ value 421for the conditional model’s parameter422(423)424γ425+426ξ427ξ428429430431432 /433 434/435* 436* 437c int438c int439int c440e441e442e443