1# RECORD b79aa4750ced5d8cf3b4aca71bb4a545f71e34bf2# TITLE: Sorting in an urban housing market—is there a response to demographic change?3# YEAR: 2021.0 VENUE: Review of Regional Research DOI: 10.1007/s10037-021-00158-74# AUTHORS: U. Neumann; Lisa Taruttis5# PAGES: 29 CANDIDATE TABLE PAGES (1-based): [12, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23]6# ABSTRACT: Using Dortmund as a case study we analyse whether rents and housing prices responded to local demographic change in a German city between 2007 and 2016. In a two-step analysis based on a spatial autoregressive hedonic pricing model and a discrete choice model of housing location we find that during the study period as a whole, higher local mortality induced a negative effect on apartment prices and rents. Yet, the neighbourhood effects of local ageing vary across sub-city districts. Most prominently, the study period was characterised by a strong and rising desire to purchase or rent housing in the vicinity of the city centre. Furthermore, prices for owner-occupied apartments and houses increased rapidly in the more well-off southern part of the city and particularly in a previously declining community, where a large-scale urban regeneration and environmental upgrading project has been implemented since 2011. The characteristics of households likely to move to this neighbourhood switched from low to high income.789===== PAGE 1 (score 5, numeric-paren 0, table-words 0, var-words 5) =====1011ORIGINAL PAPER12https://doi.org/10.1007/s10037-021-00158-713Rev Reg Res (2022) 42:111–13914Sorting in an urban housing market—is there15a response to demographic change?16Uwe Neumann117· Lisa Taruttis218Accepted: 10 September 2021 / Published online: 7 October 202119© The Author(s) 202120Abstract Using Dortmund as a case study we analyse whether rents and housing21prices responded to local demographic change in a German city between 2007 and222016. In a two-step analysis based on a spatial autoregressive hedonic pricing model23and a discrete choice model of housing location we find that during the study period24as a whole, higher local mortality induced a negative effect on apartment prices and25rents. Yet, the neighbourhood effects of local ageing vary across sub-city districts.26Availability of data and material (data transparency) This study uses data from the following27sources:281. dortmunderstatistik (2018); this data is available for download at: https://www.dortmund.de292. RWI-GEO-GRID (2018) Socio-economic data on grid level, provided by Micromarketing-Systeme30und Consult GmbH (microm). Population by age and gender, https://doi.org/10.7807/microm:31einwGeAl:v5; Share of foreigners, https://doi.org/10.7807/microm:auslaender:V5323. RWI-GEO-RED (2018) Real estate offers (purchase and rent) including object characteristics33gathered by the internet platform ImmobilienScout24. Apartments for rent, https://doi.org/10.7807/34immo:red:wm:suf:v1; Apartments for sale, https://doi.org/10.7807/immo:red:wk:v1; Houses for sale,35https://doi.org/10.7807/immo:red:hk:suf:v1364. IT.NRW Information und Technik Nordrhein-Westfalen (Statistics North Rhine-Westphalia)37(2020) Bevölkerungsstand nach Altersjahren (90) – Gemeinden – Stichtag, Fortschreibung des38Bevölkerungsstandes, 31.12.2019. 12411–09iz; this data is available for download at https://www.39landesdatenbank.nrw.de405. KOSTAT (2018) Municipal statistical data for sub-city districts, provided by KOSIS Association41KOSTAT, Bremen; this data is distributed to RWI under a contract and cannot be distributed further;42the data can be obtained from KOSTAT.43Code availability The computer code underlying the analysis was programmed using the Stata44statistical software package. The corresponding do-files are available from the authors upon request.45 Uwe Neumann46uwe.neumann@rwi-essen.de47148RWI—Leibniz Institute for Economic Research, Hohenzollernstr. 1–3, 45128 Essen, Germany49250House of Energy Markets and Finance, University of Duisburg-Essen, Universitätsstr. 12, Essen,51Germany52K535455===== PAGE 12 (score 21, numeric-paren 0, table-words 3, var-words 5) =====565712258U. Neumann, L. Taruttis59Table 2 Selected housing market characteristics (mean advertising price in Euro m2)60200861201562Apartment/ Apartment/ House/63Apartment/ Apartment/ House/64Rent65Sale66Sale67Rent68Sale69Sale70Floorspace in m271<50725.3473984.0074–756.68761017.6177–7850–70795.01801101.2281–826.11831185.5484–8570–90865.08871301.4388–895.95901443.2191–9290–120935.70941451.0595–966.76971846.5498–99>1201005.831011546.00102–1037.781041789.15105–106Year of construction107<19501085.181091039.961101318.861116.16112866.431131390.571141950–19691155.171161075.741171592.961186.111191083.101201805.101211970–19891225.341231202.781241579.221255.521261303.891271939.701281990–19991296.291301612.561311703.111326.091331767.561342066.71135Since 20001365.091371465.221382046.661397.081402036.451412193.43142Residential community143City Centre1445.701451424.951461540.711477.311481819.541492232.57150Nordstadt1514.83152797.81153858.141545.82155686.05156763.52157DO-West1584.851591095.351601668.671615.681621085.241631778.95164DO-North1654.711661143.131671576.201685.531691235.731701723.59171DO-East1725.261731262.201741869.891756.041761436.931771929.62178DO-South1795.781801571.201812062.031826.511831837.011842379.88185Hörde1864.971871055.901881295.421896.931902162.841912202.83192All1935.151941237.071951700.841966.281971440.091981764.95199Duration of offer (days),200(mean)20122.820268.820347.420427.220549.320678.4207Offers in relation to existing208stock of buildings (mean)a2090.182100.132110.112120.232130.072140.09215Observations21612,9272173027218250021912,85622027782211740222Authors’ calculation based on RWI-GEO-RED (2018)223aoffers in relation to existing buildings: 2009 and 2015224(and no German) passport has increased in all Dortmund communities, particularly225since 2012.226The population with a foreign nationality represents only part of the migrant227population, which is commonly defined as the population comprising the first and228subsequent generations of immigrants, who came to Germany since the Second229World War. While immigration has put ageing to a temporary halt in Germany as230a whole, striking local variation reveals that ageing continues to proceed rapidly in231districts with a low migrant inflow. In 2015, over 30% of the population in Dortmund232East and South were older than 60, in Nordstadt only 15% (Table 1).233As in the City Centre and Nordstadt, population growth in Hörde combined with234a rising share of foreigners (from 17.8 to 20.5%) and a slight decrease in the share235of seniors. Ageing is most advanced in the Dortmund-East and South communities,236while families with children prevail more in the North and West.237K238239240===== PAGE 13 (score 13, numeric-paren 0, table-words 3, var-words 7) =====241242Sorting in an urban housing market—is there a response to demographic change?243123244During the study period, the share of immigrants increased most rapidly in the245outer districts (especially in Dortmund-North, East and South). It is, however, still246considerably lower in the southern part of Dortmund (Dortmund-East and South)247than in the Centre and North.248From 2008 to 2015 rents in Dortmund (in current prices) increased from 5.15249up to 6.28 C per m2 of residential floorspace (+22%) on average, prices for owner-250occupied apartments from 1237 to 1440 C per m2 (+16.4%) and for one- and two-251family homes from 1701 to 1765 (+3.7%) (Table 2).252As expected by land use theory, rents and prices for housing are highest in the253City Centre. They are lowest in the West and North communities, where popula-254tion density is comparatively low. Prices are higher in the South, however, where255density is lowest among all districts. In line with a well-known North-South divide256characterising the Ruhr area due to its specific settlement history (Wehling 2003),257intra-urban price differentials reveal that the ageing southern part of the city is258nevertheless preferred as location of residence.259Rents and prices in Hörde, where a large-scale urban regeneration programme260was carried out (see above), were below the Dortmund average in 2008, i.e., before261the upgrading came into effect. By 2015, prices had increased remarkably in this262community. Rents per m2 and prices for 1-/2-familiy homes now were second only263to those in the City Centre and apartments on offer for sale cost more than in any264other community. Rents in Hörde on average increased from 4.97 to 6.93 C per m2265(+39.4%), prices for apartments from 1056 to 2163 (+105%) and for houses from2661295 to 2203 (+70%).2675 Neighbourhood stratification2685.1 Hedonic price functions for Dortmund269Estimation of hedonic price functions for rented and owner-occupied accommoda-270tion reveal considerable spatial autocorrelation among the dependent variable (as271denoted by the λ statistics in Table 3) and the residuals (cf. ρ statistics). The analysis272reveals a certain “size discount” on the rental market, as rents per m2 decrease in line273with an increase in total floorspace. No such “discount” characterises the markets274for accommodation for sale.275Regarding the characteristics of neighbourhood populations, we find that across276the complete survey period the price per m2 for apartments on offer for sale was277reduced by around 4% and rents by 1% in case ceteris paribus mortality in the278surrounding postal code district (Figs. 1 and 2) was 1% (1 death per 100 inhabitants)279higher. This may appear surprising, since prices are relatively high in the ageing280south of Dortmund. However, prices increased much more and are highest in the281“younger” City Centre and Hörde (see above). Rents and prices per m2 for apartments282were slightly lower (–0.2% and 0.6% respectively) in line with a 1% increase in the283share of foreign nationals among the residents at the level of postal codes. Whereas284local ageing corresponds to housing environments lacking the attributes of central285city quarters preferred by the younger population, apparently in Dortmund very high286K287288289===== PAGE 14 (score 97, numeric-paren 82, table-words 1, var-words 4) =====290291124292U. Neumann, L. Taruttis293Table 3 Estimates of hedonic price function, 2007–2016, GS2SLS294(1)295(2)296(3)297Rent298Purchase: Apartments299Purchase: 1-/2-Family Homes300Housing characteristics301Age categories302Yes303Yes304Yes305Age306–0.005***307–0.008***308–0.005***309(0.001)310(0.001)311(0.001)312Age23130.000***3140.000***3150.000***316(0.000)317(0.000)318(0.000)319Floor space (log)320–0.106***3210.097***3220.671***323(0.008)324(0.014)325(0.013)326No of rooms3270.014***3280.007329–0.014***330(0.003)331(0.004)332(0.002)333Multi-family housing334–0.008***335–0.018***336–337Floor level338(0.001)339(0.002)340–341Multi-family housing:342further characteristics343Yes344Yes345–346Good state of house3470.071***3480.083***3490.038***350(0.004)351(0.006)352(0.006)353Bad state of house354–0.023***355–0.105***356–0.085***357(0.004)358(0.006)359(0.005)360Further characteristics361of housing standard362Yes363Yes364Yes365Duration of offer3660.000***367–0.0003680.000369(0.000)370(0.000)371(0.000)372Annual district characteristics at 5 digits postcode level373Population (log)3740.0093750.033**3760.059***377(0.007)378(0.017)379(0.014)380Mortality (deaths per381100 inhabitants)382–0.013**383–0.036***3840.006385(0.006)386(0.011)387(0.012)388Foreigners (in %)389–0.002***390–0.006***391–0.002392(0.000)393(0.001)394(0.002)395Residential communities: time effects (reference category 2007)396Year community fixed397effects 2007–2010398Yes399Yes400Yes4012011_City Centre4020.061***4030.103***4040.276405(0.014)406(0.020)407(0.072)4082011_Nordstadt4090.003410–0.251***411–0.176**412(0.015)413(0.060)414(0.084)4152011_DO-West416–0.076***417–0.041***418–0.077419(0.012)420(0.015)421(0.064)4222011_DO-North4230.012424–0.058**425–0.116*426(0.011)427(0.027)428(0.063)4292011_DO-East4300.035***4310.007432–0.021433(0.010)434(0.013)435(0.066)436K437438439===== PAGE 15 (score 87, numeric-paren 76, table-words 1, var-words 1) =====440441Sorting in an urban housing market—is there a response to demographic change?442125443Table 3 (Continued)444(1)445(2)446(3)447Rent448Purchase: Apartments449Purchase: 1-/2-Family Homes4502011_DO-South4510.073***4520.033**4530.053454(0.013)455(0.016)456(0.065)4572011_Hörde458–0.010459–0.0324600.008461(0.016)462(0.029)463(0.073)464Year community fixed465effects 2012–2014466Yes467Yes468Yes4692015_City Centre4700.221***4710.234***4720.429***473(0.011)474(0.030)475(0.069)4762015_Nordstadt4770.146***478–0.0204790.057480(0.016)481(0.060)482(0.094)4832015_DO-West4840.132***4850.0034860.038487(0.010)488(0.022)489(0.064)4902015_DO-North4910.157***4920.050*493–0.003494(0.011)495(0.028)496(0.064)4972015_DO-East4980.147***4990.120***5000.087501(0.010)502(0.018)503(0.065)5042015_DO-South5050.166***5060.133***5070.216***508(0.012)509(0.020)510(0.065)5112015_Hörde5120.154***5130.151***5140.152**515(0.017)516(0.044)517(0.069)5182016_City Centre5190.240***5200.371***521–522(0.012)523(0.026)524–5252016_Nordstadt5260.181***5270.077528–529(0.020)530(0.063)531–5322016_DO-West5330.143***5340.075***535–536(0.011)537(0.019)538–5392016_DO-North5400.170***5410.099***542–543(0.013)544(0.029)545–5462016_DO-East5470.144***5480.159***549–550(0.012)551(0.018)552–5532016_DO-South5540.156***5550.227***556–557(0.013)558(0.019)559–5602016_Hörde5610.187***5620.267***563–564(0.019)565(0.037)566–567Further char. of plot568No569No570Yes571Year fixed effects5722008–2016573Yes574Yes575Yes576Constant5770.0085781.437***5794.707***580(0.137)581(0.529)582(0.462)583K584585586===== PAGE 16 (score 36, numeric-paren 12, table-words 7, var-words 9) =====587588126589U. Neumann, L. Taruttis590Table 3 (Continued)591(1)592(2)593(3)594Rent595Purchase: Apartments596Purchase: 1-/2-Family Homes597Observations59816,00459915,78660025,347601λ6021.271***6030.743***6040.244***605(0.062)606(0.068)607(0.036)608ρ6091.176***6101.392***6111.447***612(0.073)613(0.173)614(0.186)615Authors’ calculation based on RWI-GEO-RED (2018) (excluding observations representing the annual top6165% of offer durations per category (rent, apartments/houses for sale)), KOSTAT (2018) and dortmunder-617statistik (2018). Sample of 1-/2-family homes comprises period 2008–2015 (reference category 2008).618Accommodation on offer for rent represented by a 30% random sample. Standard errors allowing for het-619eroscedasticity in parentheses.620***/**/*= significant at 0.01/0.05/0.1-level621λ: spatial autocorrelation among dependent variable; ρ: spatial autocorrelation in error term; dependent622variable: (logarithmised) price/rent in Euro/m2 dwelling space (current prices, monthly observations, lat-623est observation per object); age categories: pre-1905, 1905–1930, 1930–1944, 1944–1960, 1960–1970,6241970–1980, 1980–1990, 1990–2000, reference category: post-2000; further characteristics of dwellings625in multi-family housing: cellar, lift, garden, balcony, fitted kitchen; further characteristics of 1-/2-family-626homes: size of plot in m2, terraced housing, building of new house is projected627shares of residents with a migrant background do not correspond to the most desired628features of urbanism.629By 2015 and 2016 rents and prices had increased in all communities. Between6302007 and 2015, apartments for sale had experienced the strongest increase in prices631per m2 in the City Centre (+23%) and in Hörde (+15%). In 2016 the upswing632accelerated even more, amounting to a 37% increase compared to 2007 in the633City Centre and to a 27% increase in Hörde. Prices increased considerably also in634Dortmund South (+22%) and East (+16%), but below 10% in the North and West635(Table 3 and Fig. 1).636It is very likely that prices in parts of Hörde had been affected by the anticipated637improvements even before new housing at the Phoenix-See entered the market. Yet,638the main increase in prices took place after 2011.639Rents increased most strongly in the City Centre also (+24%), followed by Hörde640(+18%), yet differentials across the Dortmund residential communities in the in-641crease of rents were less distinct than the increase in prices for apartments on offer642for sale. Increases in rents per m2 were lowest in the East and West communities,643but still amounted to 14% in these communities (Table 3 and Fig. 2).644Between 2008 and 2015, prices per m2 of residential floorspace increased even645more concerning houses on offer for sale in the City Centre (+43%) than for apart-646ments. Prices per m2 for houses increased significantly also in the South (+22%)647and in Hörde (+15%).6485.2 Hedonic price functions for Dortmund residential communities649In the hedonic pricing estimations for Dortmund as a whole, neighbourhood-level650price effects with respect to demographic characteristics measured at the level of651postcode zones were found to be significant concerning rented accommodation and652K653654655===== PAGE 17 (score 22, numeric-paren 13, table-words 3, var-words 3) =====656657Sorting in an urban housing market—is there a response to demographic change?658127659Fig. 1 Price change 2007–2016660per unit (Euro per m2 in cur-661rent prices) for apartments on662offer for sale in Dortmund res-663idential communities (in %);664hedonic price function (Table 3);665***/**= regression coefficients,666significant at 0.01/0.05-level,667spatial autoregressive model.668(Authors’ illustration. Map data:669GeoBasis-DE/BKG, Google.670Data sources: RWI-GEO-RED671(2018), KOSTAT (2018) and672dortmunderstatistik (2018); lines673display the boundaries of 5-digit674postal code zones)675Dortmund-West:676+7.5***677Dortmund-North:678+9.9**679Nordstadt: +7.7680City Centre: 681+37.1***682Dortmund-East: +15.9***683Dortmund-South: +22.7***684Hörde: +26.7***685Fig. 2 Price change 2007–2016686per unit (Euro per m2 in current687prices) for rented accommo-688dation in Dortmund residential689communities (in %); hedo-690nic price function (Table 3);691***/**= regression coefficients,692significant at 0.01/0.05-level,693spatial autoregressive model.694(Authors’ illustration. Map data:695GeoBasis-DE/BKG, Google.696Data sources: RWI-GEO-RED697(2018), KOSTAT (2018) and698dortmunderstatistik (2018); lines699display the boundaries of 5-digit700postal code zones)701Dortmund-West:702+14.3***703Dortmund-North:704+17.0***705Nordstadt: +18.1***706City Centre: 707+24.0***708Dortmund-East: +14.0***709Dortmund-South: +15.6***710Hörde: +18.7***711apartments and houses for sale. It is not surprising that across the complete study712period prices were higher in more densely populated districts, since apparently prices713per m2 are highest and increased most considerably in the City Centre. Further, prices714and rents were found to be reduced in line with a higher local mortality and (slightly)715by a higher share of foreigners among the local population.716As expected, hedonic price estimations for separate residential communities con-717firm that considerable spatial autocorrelation needs to be accounted for in the mea-718surement of housing prices. Regarding demographic characteristics measured at the719smaller 1km2-grid scale for each housing category, the analysis suggests that local720ageing affects prices for houses on offer for sale in a different way than prices for721apartments and rents (Table 4). In the estimations for Dortmund as a whole the coef-722ficients assigned to district-level mortality were significant and negative among rents723K724725726===== PAGE 18 (score 80, numeric-paren 62, table-words 4, var-words 1) =====727728128729U. Neumann, L. Taruttis730Table 4 Estimates of hedonic price function for residential districts, 2007–2016, GS2SLS731(1)732(2)733(3)734Rent735Purchase: Apartments7361-/2-Family Homes737Annual district characteristics at 1km grid level738Community 1 (City Centre)739Population (log)740Yes741Yes742Yes743Population aged >60744–0.0037450.0007460.027***747(in %)748(0.003)749(0.006)750(0.008)751Foreigners (in %)752–0.003**753–0.009***754–0.002755(0.001)756(0.003)757(0.005)758New stocka7590.1857600.209***7610.451762(0.159)763(0.061)764(0.644)765Observations76618397672118768670769λ7700.944***7710.585***7720.506***773ρ7740.874***7750.973***7761.079***777Community 2 (Nordstadt)778Population (log)779Yes780Yes781Yes782Population aged >607830.0017840.0037850.011786(in %)787(0.004)788(0.014)789(0.023)790Foreigners (in %)791–0.000792–0.002793–0.004794(0.002)795(0.005)796(0.008)797New stocka798–0.3607990.1588000.292801(0.254)802(0.182)803(0.204)804Observations8051149806820807404808λ8091.063***8100.674***8110.186812ρ8130.627***8140.389**8150.923***816Community 3 (Dortmund West)817Population (log)818Yes819Yes820Yes821Population aged >608220.006***8230.0008240.003825(in %)826(0.002)827(0.002)828(0.002)829Foreigners (in %)8300.003***831–0.003832–0.002833(0.001)834(0.003)835(0.002)836New stocka837–0.631***8380.0478390.156***840(0.094)841(0.118)842(0.058)843Observations844399784527678465279847λ8480.390***8490.552***8500.407***851ρ8520.939***8530.976***8541.366***855Community 4 (Dortmund North)856Population (log)857Yes858Yes859Yes860Population aged >608610.0038620.009**8630.002864(in %)865(0.002)866(0.004)867(0.003)868Foreigners (in %)869–0.0018700.001871–0.003872(0.001)873(0.002)874(0.002)875New stocka876–0.0918770.1768780.073879(0.074)880(0.167)881(0.055)882K883884885===== PAGE 19 (score 63, numeric-paren 47, table-words 4, var-words 2) =====886887Sorting in an urban housing market—is there a response to demographic change?888129889Table 4 (Continued)890(1)891(2)892(3)893Rent894Purchase: Apartments8951-/2-Family Homes896Observations89719358986878993321900λ9010.869***9020.669***9030.176*904ρ9050.660***9060.311**9071.107***908Community 5 (Dortmund East)909Population (log)910Yes911Yes912Yes913Population aged >609140.001915–0.0009160.003***917(in %)918(0.001)919(0.002)920(0.001)921Foreigners (in %)922–0.002*923–0.006**924–0.000925(0.001)926(0.003)927(0.001)928New stocka929–0.354***9300.0259310.009932(0.94)933(0.096)934(0.020)935Observations936320593732889386806939λ9400.973***9411.064***9420.171**943ρ9440.849***9450.526***9461.314***947Community 6 (Dortmund South)948Population (log)949Yes950Yes951Yes952Population aged >609530.0019540.0039550.007***956(in %)957(0.001)958(0.003)959(0.002)960Foreigners (in %)961–0.000962–0.0039630.000964(0.001)965(0.003)966(0.002)967New stocka9680.060969–0.0839700.050971(0.125)972(0.076)973(0.076)974Observations975159497624339774957978λ9790.701***9800.575***9810.437***982ρ9830.688***9840.973***9851.433***986Community 7 (Hörde)987Population (log)988Yes989Yes990Yes991Population aged >60992–0.003993–0.0129940.018995(in %)996(0.006)997(0.013)998(0.016)999Foreigners (in %)1000–0.0041001–0.01410020.0001003(0.006)1004(0.009)1005(0.008)1006and apartments for sale, but insignificant for houses (see above). In the separate es-1007timations for prices per m2 in one- or two-family houses in residential communities1008the coefficients assigned to the share of the population aged over 60 are significant1009and positive in three communities (City Centre, East and South) and insignificant in1010the other communities.1011Within the ageing East and South communities, neighbourhoods with a high share1012of seniors obviously represent surroundings deemed attractive for one- and two-1013family homes. Regarding apartments for sale, a high share of the population aged1014over 60 is assigned a positive price effect in Dortmund North, while the respective1015K101610171018===== PAGE 20 (score 23, numeric-paren 7, table-words 6, var-words 3) =====101910201301021U. Neumann, L. Taruttis1022Table 4 (Continued)1023(1)1024(2)1025(3)1026Rent1027Purchase: Apartments10281-/2-Family Homes1029New offersa1030–0.1431031–0.3771032–0.7961033(0.226)1034(0.278)1035(0.574)1036Observations1037626103864610394751040λ10410.760***10420.518***10430.1821044ρ10450.01710460.740***10470.918***1048Housing characteristics1049Yes1050Yes1051Yes1052Year fixed effects1053Yes1054Yes1055Yes1056Constants1057Yes1058Yes1059Yes1060Authors’ calculation based on RWI-GEO-RED (2018) (excluding observations representing the annual1061top 5% of offer durations per category (rent, apartments/houses for sale)) and RWI-GEO-GRID (2018).1062Sample of 1-/2-family homes comprises period 2008–2015 (reference category 2008). Accommodation on1063offer for rent is represented by a 30% random sample. Standard errors allowing for heteroscedasticity in1064parentheses1065***/**/*= significant at 0.01/0.05/0.1-level1066Dependent variable: (logarithmised) price/rent in Euro/m2 dwelling space (current prices, monthly obser-1067vations, latest observation per object); housing characteristics cf. Table 31068aAnnual offerings in relation to the existing stock of residential buildings1069coefficients turn out insignificant in all other districts. Rents are affected (positively)1070only in Dortmund West (Table 4).1071Rents and prices for apartments were affected (slightly) negatively by higher local1072shares of foreigners in the City Centre and Dortmund East, but rents were affected1073(slightly) positively by a higher share of foreigners among the local neighbourhoods1074of Dortmund West. Whereas local ageing and migrant agglomeration at the level1075of larger districts (representing around 20,000 inhabitants on average) tend to be1076regarded as disadvantageous, at the level of 1km2 grids both characteristics may1077affect local housing environments in the opposite direction. A larger share of senior1078citizens, for example, may be attributed to a safe neighbourhood at the very local1079level. Within a suburban district, on the other hand, a comparatively higher share of1080migrants may correspond to more central locations where prices are higher.1081In the City Centre and in the East and West communities, housing prices appar-1082ently were affected to a significant extent by the volume of housing units entering1083the market in a specific year. In the Centre, average prices for apartments were even1084higher in case a larger number of apartments were on offer for sale in the respec-1085tive neighbourhood. Rents were affected in the East and West communities, where1086a large local volume of rental offers coincided with a price reduction.10875.3 Community-level variation in location choice probabilities1088The third step of our analysis examines the likelihood of households to have chosen1089accommodation in one of the seven Dortmund residential communities in each year1090during the period from 2009 to 2016.1091In 2009, the mean sizes of households likely to move into rented accommo-1092dation were smaller than in Dortmund-West (which was the base category) in all1093K109410951096===== PAGE 21 (score 120, numeric-paren 98, table-words 5, var-words 1) =====10971098Sorting in an urban housing market—is there a response to demographic change?10991311100Table 5 Location choice probabilities, by Dortmund residential community (Multinomial logit estimation, selected results for 2009 and 2015/2016)1101City Centre1102Nordstadt1103DO-West1104DO-North1105DO-East1106DO-South1107Hörde110820091109Rent1110Est. household size1111–1.106***1112–0.458***1113–11141.074***11150.0231116–0.459***1117–0.693***1118(log)1119(0.060)1120(0.072)1121–1122(0.071)1123(0.058)1124(0.067)1125(0.110)1126Total rent/pp (log)11271.164***11281.144***1129–1130–0.892***1131–0.07011320.398***11330.803***1134(0.062)1135(0.070)1136–1137(0.069)1138(0.059)1139(0.069)1140(0.103)1141a0h;t1142–0.396***1143–0.910***1144–1145–0.718***1146–0.0371147–0.663***1148–1.817***1149(0.036)1150(0.043)1151–1152(0.041)1153(0.032)1154(0.038)1155(0.060)1156Observations115795241158Pseudo-R211590.0461160Apartments for sale1161Est. household size1162–11632.038***11640.634***11652.136***11660.621***11670.302***11680.1951169(log)1170–1171(0.102)1172(0.089)1173(0.132)1174(0.091)1175(0.087)1176(0.150)1177Total price/pp (log)1178–1179–2.685***1180–0.820***1181–1.871***1182–0.706***1183–0.0311184–0.589***1185–1186(0.115)1187(0.102)1188(0.137)1189(0.103)1190(0.090)1191(0.176)1192a0h;t1193–1194–0.667***1195–0.0021196–1.513***1197–0.0551198–0.501***1199–1.605***1200–1201(0.070)1202(0.055)1203(0.096)1204(0.056)1205(0.063)1206(0.096)1207Observations120836561209Pseudo-R212100.0821211Houses for sale1212Est. household size12130.589***12141.841***1215–12160.671***12170.1491218–0.116***12191.066***1220(log)1221(0.131)1222(0.160)1223–1224(0.098)1225(0.093)1226(0.098)1227(0.154)1228Total price/pp (log)1229–0.1371230–0.967***1231–1232–0.911***1233–0.677***1234–0.195***1235–0.683***1236(0.137)1237(0.164)1238–1239(0.098)1240(0.086)1241(0.089)1242(0.172)1243a0h;t1244–1.767***1245–3.617***1246–1247–0.674***1248–0.0071249–0.155**1250–2.608***1251(0.117)1252(0.273)1253–1254(0.080)1255(0.063)1256(0.065)1257(0.172)1258K125912601261===== PAGE 22 (score 102, numeric-paren 79, table-words 7, var-words 1) =====126212631321264U. Neumann, L. Taruttis1265Table 5 (Continued)1266City Centre1267Nordstadt1268DO-West1269DO-North1270DO-East1271DO-South1272Hörde1273Observations127421141275Pseudo-R212760.057127720161278Rent1279Est. household size1280–1.955***12810.868***1282–12831555***1284–0.790***1285–0.609***1286–0.619***1287(log)1288(0.055)1289(0.0965)1290–1291(0.072)1292(0.048)1293(0.053)1294(0.062)1295Total rent/pp (log)12963.005***1297–0.212***1298–1299–1.550***13001.333***13011.034***13021.864***1303(0.064)1304(0.085)1305–1306(0.082)1307(0.055)1308(0.060)1309(0.069)1310a0h;t1311–0.570***1312–1.631***1313–1314–1.892***1315–0.0441316–0.405***1317–0.983***1318(0.041)1319(0.060)1320–1321(0.067)1322(0.033)1323(0.0036)1324(0.044)1325Observations132611,1431327Pseudo-R213280.1441329Apartments for sale1330Est. household size1331–0.417***13322.509***13330.921***13342.352***1335–1336–0.0411337–0.2061338(log)1339(0.115)1340(0.142)1341(0.103)1342(0.152)1343–1344(0.113)1345(0.140)1346Total price/pp (log)13471.080***1348–4.092***1349–1.370***1350–2.705***1351–13520.490***13532.009***1354(0.133)1355(0.193)1356(0.124)1357(0.186)1358–1359(0.129)1360(0.173)1361a0h;t1362–1.114***1363–2.349***1364–0.252***1365–1.942***1366–1367–0.866***1368–2.249***1369(0.089)1370(0.145)1371(0.064)1372(0.125)1373–1374(0.079)1375(0.151)1376Observations137725121378Pseudo-R213790.1961380Houses for sale (2015)1381Est. household size1382–0.25513833.102***1384–13850.274***1386–0.593***1387–0.668***1388–0.509**1389(log)1390(0.186)1391(0.394)1392–1393(0.092)1394(0.100)1395(0.110)1396(0.226)1397K139813991400===== PAGE 23 (score 34, numeric-paren 23, table-words 4, var-words 1) =====14011402Sorting in an urban housing market—is there a response to demographic change?14031331404Table 5 (Continued)1405City Centre1406Nordstadt1407DO-West1408DO-North1409DO-East1410DO-South1411Hörde1412Total price/pp (log)14131.764***1414–0.816***1415–1416–0.437***14170.767***14181.161***14191.262***1420(0.181)1421(0.256)1422–1423(0.105)1424(0.103)1425(0.113)1426(0.219)1427a0h;t1428–2.600***1429–6.213***1430–1431–0.343***1432–0.0541433–0.422***1434–2.477***1435(0.197)1436(0.757)1437–1438(0.080)1439(0.072)1440(0.081)1441(0.187)1442Observations144316511444Pseudo-R214450.1071446Authors’ calculation based on RWI-GEO-RED (2018), excluding observations representing the annual top 5% of offer durations per category (rent, apartments/houses for1447sale). Robust standard errors in parentheses1448***/**/*= significant at 0.01/0.05/0.1-level1449est. household size= floorspace of dwelling in m2, divided by the average floorspace per person in 2013 in corresponding residential community (Stadt Dortmund 2017);1450total rent/pp: total rent, divided by the average monthly purchasing power per household at the level of 1km grids; total price/pp: total price, divided by the average annual1451purchasing power per household; a0h;t Dannual choice probability among households with mean characteristics1452K145314541455===== PAGE 24 (score 9, numeric-paren 0, table-words 3, var-words 3) =====145614571341458U. Neumann, L. Taruttis1459districts except for Dortmund North and East. Total rents in relation to household1460income were higher in all districts than in Dortmund-West, except for North and1461East (Table 5).1462In general, considering housing on offer for rent or sale mean and household-1463specific probabilities characterising the likelihood to choose among six out of seven1464Dortmund residential communities remained relatively stable over the study period,1465e.g., one- and two-person-households being likely to move to the City Centre and1466low-income households prone to choose Nordstadt. A considerable change in the1467choice probabilities relating to household characteristics, however, took place with1468respect to Hörde. This change corresponds to a considerable upswing in prices for1469apartments and houses on offer for sale, which is likely to result at least partly1470from the environmental upgrading due to the Emscher Conversion and Phoenix-1471See development. In 2009, households moving into purchased accommodation in1472Hörde were likely to earn a comparatively low income (as denoted by negative1473coefficients assigned to the total price of the apartment or house in relation to local1474income). In 2016, they were likely to be well-off and—in case they were moving1475into apartments—to pay higher total prices even than households moving to the City1476Centre (Table 5).1477For households with mean characteristics (as denoted by the coefficients for1478a0h,t), however, it remained improbable to choose accommodation for rent or sale in1479Hörde. Regarding rented accommodation and one- or two-family homes for sale, the1480Dortmund West community represented mean preferences most accurately in 20091481and 2015/16. With respect to owner-occupied apartments, mean preferences shifted1482from the City Centre (2009) to the Dortmund East community. Whereas total prices1483in relation to local income differed only insignificantly between the Centre and East1484in 2009, in 2016 total prices for apartments in the Centre had become considerably1485(i.e. over one standard deviation, Table 5) higher than in the East.1486While both the Centre and Hörde experienced an upswing of local housing prices,1487Nordstadt was affected by a continuing downturn, specifically concerning owner-1488occupied apartments. While the size of households likely to move in remained1489comparatively large, total prices remained considerably below those in the other1490districts and the gap between Nordstadt and the other districts in this respect has1491widened. In 2009, Nordstadt prices in relation to local income were 2.69 standard1492deviations below those in the reference category (which was the City Centre), in14932016 they were over 4 standard deviations below the reference category (Dortmund1494East).14956 Conclusions1496We analyse whether housing prices and location choice probabilities in Dortmund1497over the period from 2007 to 2016 indicate a response to local demographic change1498at the level of different neighbourhood scales. It is difficult to isolate the effects of1499local ageing from other determinants of residential sorting in empirical analysis. We1500therefore apply different methods and incorporate neighbourhood-level statistics at1501varying spatial scales in order to evaluate the robustness of our results.1502K1503