1# RECORD cc4e9f217dbac07b4b73c8b633af8dfec0f75eb12# TITLE: Property price and anti-social behaviour: An empirical study of Northern Ireland3# YEAR: 2025.0 VENUE: International Journal of Strategic Property Management DOI: 10.3846/ijspm.2025.239834# AUTHORS: D. Lo; Yao-Xuan Huang; Michael McCord; Y. Yau; John McCord; Tariq Aziz; M. Haran; Lara Al-haddad5# PAGES: 17 CANDIDATE TABLE PAGES (1-based): [7, 8, 9, 10, 11, 12]6# ABSTRACT: This study examines the impact of anti-social behaviour (ASB) on property prices. By analysing over 14,500 market transactions in Northern Ireland, we find that the prevalence of ASB within a neighbourhood exerts a direct and negative influence on house prices, albeit with diminishing effect at the margin. Furthermore, the dampening effects of ASB are more pronounced in districts characterised by higher population density, proximity to the capital city (Belfast), and lower property values. A thorough analysis of district-level data across a wide range of statistical indicators further indicates that the adverse impact of ASB on property prices is most acute in areas marked by social and economic deprivation, including factors such as income, employment, education, and access to services. Lastly, our submarket analysis suggests that the apartment sector and public housing are disproportionately affected by ASB in terms of price depreciation compared to other property types.789===== PAGE 1 (score 13, numeric-paren 0, table-words 0, var-words 13) =====1011This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), 12which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.13Copyright © 2025 The Author(s). Published by Vilnius Gediminas Technical University14ISSN: 1648-715X / eISSN: 1648-917915INTERNATIONAL JOURNAL 16OF STRATEGIC PROPERTY 17MANAGEMENT18202519Volume 2920Issue 621Pages 469–48522https://doi.org/10.3846/ijspm.2025.2398323PROPERTY PRICE AND ANTI-SOCIAL BEHAVIOUR: AN EMPIRICAL STUDY OF 24NORTHERN IRELAND25Daniel LO 1*, Yaoxuan HUANG 2, Michael McCORD 1, Yung YAU 3, John McCORD 4, 26Tariq AZIZ 5, Martin HARAN 1, Lara AL-HADDAD 6271 Faculty of Computing, Engineering and the Built Environment, Ulster University, Belfast, the United Kingdom282 School of Graduate Studies, Lingnan University, Hong Kong, China293 School of Graduate Studies and Department of Sociology and Social Policy, Lingnan University, Hong Kong, China304 Faculty of Law, Ulster University, Belfast, the United Kingdom315 College of Business Administration, Prince Muhammad Bin Fahd University, Dhahran, Saudi Arabia326 Faculty of Business, Yarmouk University, Irbid, Hashemite Kingdom of Jordan33Article History: 34Abstract. This study examines the impact of anti-social behaviour (ASB) on property prices. By analysing over 3514,500 market transactions in Northern Ireland, we find that the prevalence of ASB within a neighbourhood 36exerts a direct and negative influence on house prices, albeit with diminishing effect at the margin. Further-37more, the dampening effects of ASB are more pronounced in districts characterised by higher population 38density, proximity to the capital city (Belfast), and lower property values. A thorough analysis of district-level 39data across a wide range of statistical indicators further indicates that the adverse impact of ASB on property 40prices is most acute in areas marked by social and economic deprivation, including factors such as income, 41employment, education, and access to services. Lastly, our submarket analysis suggests that the apartment 42sector and public housing are disproportionately affected by ASB in terms of price depreciation compared to 43other property types.44 ■received 23 February 202545 ■accepted 3 April 202546Keywords: property price, anti-social behaviour, hedonic modelling, Northern Ireland, housing market.47 * Corresponding author. E-mail: d.lo@ulster.ac.uk481. Introduction49Anti-social behaviour (ASB) poses a significant challenge 50in built environments, particularly within urban residential 51property markets, where factors such as high population 52density, the close proximity of dwellings, diverse socioeco-53nomic demographics, and frequent tenant turnover inten-54sify its visibility and impact. Broadly defined as actions that 55cause harassment, alarm, or distress to others, ASB encom-56passes a wide array of activities, from noise disturbances 57to vandalism and violence (Anti-social Behaviour, Crime 58and Policing Act 2014, the U.K.). The presence of ASB in a 59neighbourhood can lead to declining property values, cre-60ate an inhospitable living environment, and deter poten-61tial buyers and traders, thereby exerting a negative impact 62on the real estate market. Studies have shown that areas 63with elevated levels of ASB generally experience stunted 64community development as prospective homebuyers and 65investors are dissuaded by the perceived risks associated 66with such neighborhoods, including diminished quality of 67life and safety concerns (Bannister & Kearns, 2013). Con-68sequently, understanding the dynamics of ASB is essential 69for real estate stakeholders, including homebuyers, prop-70erty developers, and urban planners, as it plays a pivotal 71role in shaping housing supply and demand as well as 72market performance.73In the U.K., ASB has developed into a notable soci-74etal issue that cannot be disregarded, particularly within 75the urban centres of major cities with high population 76density. Although the Anti-social Behaviour Act 2003 and 77subsequent legislative measures have been implemented 78to mitigate the detrimental effects of ASB on communi-79ties, its persistence continues to challenge policymakers 80(Home Office, 2014). Despite these legislative efforts, the 81enduring prevalence of ASB underscores the difficulties in 82effectively curbing its impact, rendering it a pervasive and 83complex problem for those tasked with developing and 84enforcing policy (Home Office, 2014). The linkage between 85ASB and real estate markets can indeed be understood 86through the concept of “stigma” associated with certain 87neighbourhoods. As Gibbons (2004) observed, areas bur-88dened with high levels of crime and ASB often develop 89a negative reputation, resulting in reduced demand for 90housing and, consequently, shrinking property values. For 919293===== PAGE 7 (score 18, numeric-paren 0, table-words 2, var-words 10) =====9495International Journal of Strategic Property Management, 2025, 29(6), 469–48596475974. Data and descriptive statistics98The sample data for residential market transactions was 99obtained from various local real estate agencies in North-100ern Ireland, covering the period from the first quarter of 1012018 to the second quarter of 2022. This dataset includes 102sales prices alongside detailed property-level information 103such as building age, property size, address, property type, 104Energy Performance Certificate (EPC) scores and building 105conditions. This data was then merged with a dataset 106from the Northern Ireland Statistical and Research Agency, 107which provides Super Output Area (SOA)-level information 108on ASB and socioeconomic characteristics, as outlined in 109Section 3. In total, approximately 14,500 transaction obser-110vations were utilised for the Hedonic regression modelling.111Table 2. Composition of the sample 112Property type113Detached114Semi-detached115Terraced116Apartment117Townhouse1184018119(29.28%)1204990121(36.36%)1223571123(26.02%)124520125(3.79%)126626127(4.56%)128Belfast129Within Belfast130Outside Belfast1316231132(42.91%)1338290134(57.09%)135Private vs public136Private137Public13812,200139(84.02%)1402321141(15.98%)142Energy Performance 143Certificate144A145B146C147D148E14911150(0.08%)151335152(2.31%)15311,943154(82.26%)1552,226156(15.33%)1574158(0.03%)159Geography 160Rural161Urban1621649163(11.76%)16412377165(88.24%)166Table 3. Descriptive statistics167Aggregate168Variable169Average170Median171Q1172Q3173Stand. Dev.174Property price (£)175193,900176162,000177119,500178230,000179145,124180ASB (per 1000 pop.)18126.5118221.1018314.2218433.1018519.98186Submarket (property type)187Detached188Property price (£)189292,394190250,000191190,000192345,000193194,025194ASB (per 1000 pop.)19518.6919615.0019710.9019822.1019913.07200Semi-detached201Property price (£)202173,908203149,940204131,000205199,95020689,880207ASB (per 1000 pop.)20823.8720919.721014.0021131.0021215.56213Terraced214Property price (£)215116,392216104,00021781,000218135,000219104,795220ASB (per 1000 pop.)22139.1522232.622323.22244622525.95226Apartment227Property price (£)228143,613229129,25023094,950231169,95023292,866233ASB (per 1000 pop.)23431.3923525.42361623740.323823.06239Townhouse240Property price (£)241176,391242153,000243125,950244208,00024592,286246ASB (per 1000 pop.)24725.9924820.724914.825031.825117.72252Belfast253Property price (£)254192,864255152,000256110,500257230,000258140,701259ASB (per 1000 pop.)26030.5926125.62621626339.726420.87265Non-Belfast266Property price (£)267196,433268167,000269125,000270231,500271148,3666272ASB (per 1000 pop.)27324.0727418.62751327630.427719.02278Private279Property price (£)280205,757281167,500282125,000283247,500284143,315285ASB (per 1000 pop.)28624.6628720.228812.228930.129020.85291Public292Property price (£)293185,144294159,950295115,000296220,000297148,490298ASB (per 1000 pop.)29926.1830022.130113.730232.630318.01304305306===== PAGE 8 (score 41, numeric-paren 0, table-words 10, var-words 17) =====307308476309D. Lo et al. Property price and anti-social behaviour: An empirical study of Northern Ireland310Table 2 summarises the composition of the dataset. 311The majority of the observations are either detached 312(29.28%), semi-detached (36.36%), or terraced (26.02%) 313properties, with apartments and townhouses accounting 314for only 3.79% and 4.56% of the sample, respectively. In 315addition, 88.24% of the properties are located in urban 316areas, with 42.91% situated within the Belfast local gov-317ernment district. Notably, a significant proportion of the 318properties (84.02%) were privately built. Regarding EPC 319scores, 82.26% of the observations achieved an average 320energy efficiency rating of C, with an additional 15.33% 321receiving a D rating.322Table 3 provides descriptive statistics of the sample 323data at both aggregate and submarket levels, focusing on 324ASB intensity and property prices. The average and me-325dian house prices in Northern Ireland are approximately 326£193,900 and £162,000, respectively, indicating a market 327skewed towards lower-priced housing. The average ASB 328intensity is 26.51 incidents per 1,000 population per year, 329with the top 25% most ASB-affected SOAs exceeding 33033.10 incidents.331Submarket-level analysis reveals a strong negative cor-332relation between property prices and ASB. For instance, 333terraced housing exhibits the highest average ASB intensi-334ty (39.15 incidents) and the lowest average property price 335(£116,392). Conversely, detached houses, which command 336the highest average price (£292,394), are the least affected 337by ASB, with an average intensity of 18.69 incidents. Other 338property types fall between these two extremes in terms 339of both ASB intensity and property prices.340Furthermore, the data shows a notable difference be-341tween SOAs in Belfast and those in other regions, with 342Belfast averaging 30.59 ASB incidents compared to 24.07 343in other areas, despite generally higher house prices in 344Belfast. Publicly-built residential properties also tend to 345experience higher ASB rates compared to privately-built 346ones. The average and median ASB figures for publicly-347built properties are 26.18 and 22.1, respectively, whereas 348the corresponding figures for privately-built properties are 34924.66 and 20.2.350Our correlation analysis, presented in Table 4, fur-351ther supports the negative relationship between property 352prices and ASB, as evidenced by a correlation coefficient 353of –0.24. This finding provides initial evidence that ASB 354activity within a neighbourhood may depress property val-355ues. Additionally, the analysis indicates a positive correla-356tion between ASB and population density, suggesting that 357ASB is more prevalent in overcrowded neighbourhoods. 358Finally, all socioeconomic ranks, except for accessibility, 359show a negative correlation with ASB, indicating that ar-360eas affected by ASB generally have poorer socioeconomic 361conditions.3625. Empirical findings and discussion363The findings from the fixed effects hedonic analysis is 364presented in three parts. First we examine the effect of 365ASB at the general level accounting for housing attributes 366and locational dummy variables and develop a series of 367interaction terms to further account for, isolate and de-368termine whether urban population density, property value 369and public housing have an impact upon the nature and 370level of the effect (Models 1–6). Secondly, we develop a 371series of property type models (Models 7–11), to examine 372ASB and scrutinise and distill whether there is a differen-373tial effect on property prices. Lastly, we specify a number 374socioeconomic ranking models which are based on the 375ranking of multiple deprivation measure (Model 12) and 376the underpinning domains of deprivation such as income, 377employment, education, access and crime (Models 13–17) 378to establish whether deprivation impacts on house prices 379with respect to ASB. Indeed, this approach aligns with Mc-380Cann (2013), who suggested that socio-economic factors, 381such as income and employment levels, influence residen-382tial preferences and the value placed on neighbourhoods 383affected by ASB.384Overall, the initial hedonic base models (Models 1 and 2) 385and property attribute models (Models 3–6) in Table 5 386show adjusted R2 values ranging between 63.1% and 38782.3% indicating the coefficients to explain between 63% 388and 82% in the variation of sales price, with the coefficients 389displaying the expected sign, statistical significance and 390reasonable level of magnitudes. For instance, an increase 391in property size increases the price of the dwelling with 392one additional square metre of living space increases the 393price by between 0.51% and 0.84% (Models 1–6). Similarly, 394the age of the property comprises a value significant effect 395ranging between 0.009% and 0.057%, and if the property 396is privately built this exhibits a positive effect ranging be-397tween 5.9% and 16.8% across the models (Models 1–6). 398Inspection of the rural coefficient also reveals a negative 399and statistically significant effect ranging between 3.99% 400and 8.75%, inferring that rural properties are lower in value 401relative to their urban counterparts. This finding resonates 402with the work of Braakmann (2012), who observed that 403urban areas tend to experience a greater depreciation in 404property values due to factors like social disorder, a con-405clusion that is particularly relevant in the context of ASB.406Turning to the parameter of interest, namely ASB, the 407base models (Models 1 and 2) reveal that ASB tends to de-408press house prices, demonstrating that when the number 409Table 4. Correlation analysis410Price411ASB412ASB413–0.239764141.00000415Density416–0.237664170.372488418Income rank4190.269692420–0.33513421Employment rank4220.406457423–0.44976424Education rank4250.482588426–0.46102427Accessibility rank428–0.139404290.390209430Crime rank4310.271278432–0.59419433MDM rank4340.385611435–0.58832436437438===== PAGE 9 (score 73, numeric-paren 44, table-words 4, var-words 7) =====439440International Journal of Strategic Property Management, 2025, 29(6), 469–485441477442action term with ASB (Model 3), the results suggest that 443neighbourhoods of a higher living density tend to have 444a stronger impact of ASB on house prices, i.e., property 445values in more densely populated neighbourhoods are 446more affected by the levels of ASB. In a similar vein, the 447interaction term between price and ASB (Model 4) exhibit 448properties of lower value to be more affected by ASB, with 449a more pronounced negative price effect stemming from 450ASB being observed in public estates/properties than in 451the private counterparts (Model 5). In addition, the results 452show that in terms of property pricing, the capital city of 453Northern Ireland, Belfast, is more affected by ASB as dis-454played by the negative and statistically significant coef-455ficient on ASB × Belfast (Model 6). This finding aligns with 456Besley and Mueller (2012), who highlighted that the effects 457of ASB cases per 1,000 population increases by 10%, prop-458erty prices tend to decrease between 0.39% and 0.95% 459depending on whether marginal effects are accounted for. 460This result is consistent with the findings of Besley and 461Mueller (2012), who found that violence and other socially 462undesirable effects in Belfast resulted in a reduction of 463house prices by 8–14%. As observed in Model 2, a dimin-464ishing marginal effect of ASB on price is also detected as 465indicated by the positive sign of the squared term on ASB. 466This diminishing effect suggests that, as the level of ASB 467rises, the impact on prices becomes less pronounced, simi-468lar to the results of Seo (2018), who found that neighbour-469hood disorder showed decreasing returns in its influence 470on housing prices as disorder increased. When further ac-471counting for urban (neighbourhood) density and its inter-472Table 5. Housing attribute models473Model 1474(Base)475Model 2476(Base 2)477Model 3478(Density)479Model 4480(Property price)481Model 5482(Private)483Model 6484(Belfast)485Constant48610.79874487(0.132176)***48811.18415489(0.130570)***49011.18743491(0.130277)***49211.39189493(0.092940)***49411.31744495(0.128098)***49611.17922497(0.130478)***498ASB499–0.003930500(0.000164)***501–0.009514502(0.000366)***503–0.008774504(0.000376)***505–0.182533506(0.001480)***507–0.009762508(0.000455)***509–0.008565510(0.000417)***511ASB × ASB5124.83×10–5513(2.81×10–6)***5145.22×10–5515(2.84×10–6)***5164.48×10–5517(2.00×10–6)***5184.86×10–5519(2.84×10–6)***5204.61×10–5521(2.85×10–6)***522Age5230.000901524(0.000118)***5250.002997526(0.000507)***5270.002931528(0.000506)***5290.002727530(0.000361)***5310.002858532(0.000507)***5330.002771534(0.000509)***535Age × Age536–1.42×10–5537(3.80×10–6)***538–1.30×10–5539(3.80×10–6)***540–1.47×10–5541(2.71×10–6)***542–1.28 ×10–5543(3.80×10–6)***544–1.28×10–5545(3.81×10–6)***546Size5470.005028548(6.64×10–5)***5490.008369550(0.000173)***5510.008245552(0.000173)***5530.005017554(0.000126)***5550.008363556(0.000173)***5570.008346558(0.000173)***559Size × Size560–8.31×10–6561(3.90×10–7)***562–8.12×10–6563(3.90×10–7)***564–3.92×10–6565(2.80×10–7)***566–8.31×10–6567(3.90×10–7)***568–8.27×10–6569(3.90×10–7)***570Parking onsite5710.085923572(0.010863)***5730.063596574(0.010765)***5750.054154576(0.010803)***5770.027686578(0.007667)***5790.064033580(0.010774)***5810.060777582(0.010773)***583Rural584–0.071657585(0.010610)***586–0.087541587(0.010440)***588–0.088568589(0.010417)***590–0.039933591(0.007440)***592–0.088370593(0.010449)***594–0.079871595(0.010557)***596Private5970.138356598(0.024704)***5990.124989600(0.024144)***6010.134252602(0.024116)***6030.059928604(0.017191)***6050.168379606(0.028833)***6070.124831608(0.024126)***609ASB × Density610–3.74×10–5611(4.60×10–6)***612ASB × Price6130.014814614(0.000125)***615ASB × Private6160.001227617(0.000446)***618ASB × Belfast 619–0.001432620(0.000302)***621Time Dummies622Yes623Yes624Yes625Yes626Yes627Yes628LGD Dummies629Yes630Yes631Yes632Yes633Yes634Yes635Property Type 636Dummies637Yes638Yes639Yes640Yes641Yes642Yes643EPC Dummies644Yes645Yes646Yes647Yes648Yes649Yes650Ext Dummies651Yes652Yes653Yes654Yes655Yes656Yes657Obs65814,52165914,52166014,52166114,52166214,52166314,521664R2 6650.6320066660.6513386670.6529276680.8234216690.6515206700.651877671Adj R26720.6308626730.6501816740.6517526750.8228236760.6503406770.650698678F679552.4417680563.2336681555.57916821377.164683552.1440684553.0134685Prob(F)6860.0000006870.0000006880.0000006890.0000006900.0000006910.000000692Notes: LGD = Local government district; EPC = Energy Performance Certificate; Ext = Condition of external repair. *** indicates 1% statistical significance. 693694695===== PAGE 10 (score 77, numeric-paren 34, table-words 6, var-words 19) =====696697478698D. Lo et al. Property price and anti-social behaviour: An empirical study of Northern Ireland699of ASB are more pronounced in urban centres such as 700Belfast, where population density and social heterogeneity 701likely exacerbate the effects of ASB.702Table 6 exhibits the role of ASB by property type (Mod-703els 7–11). The findings reveal that amongst all property 704types, apartments seemingly are the most affected by ASB 705(–1.307, p < .001), followed by semi-detached, terraced, 706townhouse and detached with an ASB-induced property 707price discount ranging between 0.66% to 1.13%. This find-708ing suggests that this may be largely due to the effect 709of density, the concentration of younger people gather-710ing in more urbanized areas or “hot spots” of highly con-711centrated space and the potential inter-heterogeneity in 712areas or surrounding areas and how the neighbourhood 713structure is defined. This is concomitant with the findings 714of Ceccato and Wilhelmsson (2011) who observed that ad-715jacency to public places can impact upon, and reveal signs 716of physical and social disorder in which environmental 717cues translate into weakened social ties among residents, 718a negative demand for housing in an area and a nega-719tive effect on property prices. Indeed, of particular note, 720Ceccato and Wilhelmsson (2011) observed vandalism to 721have a significant and independent effect on apartment 722prices in Stockholm municipality indicating that the effects 723of crime vary by housing type and have differential effects 724on house prices. 725Examination of the role of deprivation and its domains 726and the effects of ASB on house prices are presented in 727Table 7. As observed in Model 12 which takes into ac-728count the headline measure of multiple deprivation rank-729ing, the effects of ASB on house price are more significant 730in neighbourhoods of higher multiple deprivation scores. 731This finding is in line with Knox (2011), who argued that 732higher levels of deprivation correlate with increased ASB, 733and suggests that deprived areas are more vulnerable 734to the negative price effects of ASB. Further isolating 735the effects of ASB on house prices applying the sepa-736rate domains of deprivation (Models 13–17) reveals that 737the effects of ASB on house price are more significant 738in lower-income neighbourhoods and in neighbourhoods 739comprising higher unemployment rates. Moreover, the re-740sults show the effects of ASB on house price are more 741significant in neighbourhoods of lower educational attain-742ment, reduced accessibility to services and in neighbour-743hoods which constitute higher crime rates. These results 744align with Haydon and Scraton (2000), who noted that 745socio-economic factors, such as poverty and inequality, 746exacerbate ASB in Northern Ireland, contributing to its 747Table 6. Property type models748Model 7749(Detached)750Model 8751(Semi Detached)752Model 9753(Terraced)754Model 10755(Apartment)756Model 11757(Townhouse)758Constant75911.43231760(0.218770)***76110.48241762(0.191199)***76310.95269764(0.156221)***76511.50655766(0.223572)***76710.53464768(0.201467)***769ASB770–0.006672771(0.000834)***772–0.011330773(0.000732)***774–0.007973775(0.000642)***776–0.013072777(0.002463)***778–0.006937779(0.001467)***780ASB × ASB7812.65×10–5782(7.97×10–6)***7836.88×10–5784(7.21×10–6)***7854.01×10–5786(4.21×10–6)***7870.000104788(2.39×10–5)***7891.80×10–5790(1.28×10–5)791Age7920.006723793(0.000856)***7940.000851795(0.000923)796–0.002881797(0.001315)**798–0.005064799(0.002595)**8000.003915801(0.002399)*802Age × Age803–3.41×10–5804(6.40×10–6)***8059.60×10–6806(7.21×10–6)8071.90×10–5808(9.28×10–6)**8096.06×10–5810(1.95×10–5)***811–2.86×10–5812(1.86×10–5)813Size8140.007221815(0.000264)***8160.010776817(0.000460)***8180.008178819(0.000864)***8200.007011821(0.001226)***8220.011030823(0.000841)***824Size × Size825–6.12×10–6826(5.23×10–7)***827–1.40×10–5828(1.28×10–6)***829–5.60×10–6830(3.44×10–6)*831–1.96×10–5832(3.97×10–6)***833–1.50×10–5834(2.39×10–6)*** 835Parking onsite836–0.011208837(0.043728)8380.063166839(0.030378)**8400.027108841(0.014338)*842–8430.091669844(0.029170)***845Rural846–0.079585847(0.015479)***848–0.107333849(0.021478)***850–0.050162851(0.027596)*8520.044664853(0.083693)854–0.019899855(0.038885)856Private857–0.008144858(0.089321)8590.231046860(0.046031)***8610.070744862(0.036865)**8630.275627864(0.076130)***8650.277295866(0.067088)***867Time Dummies868Yes869Yes870Yes871Yes872Yes873LGD Dummies874Yes875Yes876Yes877Yes878Yes879EPC Dummies880Yes881Yes882Yes883Yes884Yes885Ext Dummies886Yes887Yes888Yes889Yes890Yes891Obs8924,0188934,9908943,571895520896626897R28980.5233328990.4714489000.3797249010.4929089020.790389903Adj R29040.5180539050.4669609060.3723409070.4617989080.777955909F91099.13489911105.060291251.4236891315.8440691463.56403915Prob(F)9160.0000009170.0000009180.0000009190.0000009200.000000921Notes: LGD = Local government district; EPC = Energy Performance Certificate; Ext = Condition of external repair. *, **, *** indicate 10%, 5% and 1% statistical 922significance respectively.923924925===== PAGE 11 (score 81, numeric-paren 41, table-words 8, var-words 9) =====926927International Journal of Strategic Property Management, 2025, 29(6), 469–485928479929more severe effects on property values. Viewed from an-930other perspective, high-income households appear will-931ing to pay a price premium to distance themselves from 932areas affected by anti-social behaviour and its associated 933activities. This observation aligns with the conceptual ar-934gument put forth by McCann (2013), who suggests that 935income-based residential segregation emerges as affluent 936individuals increasingly cluster within certain neighbour-937hoods, driven not only by considerations such as com-938muting costs and proximity to employment centres but 939also by the desire to secure a safer and more desirable 940living environment.941Robustness check3942As a robustness check for potential spatial autocorre-943lation in property prices, as well as omitted variables that 944could bias our results, we developed three spatial lagged 9453 We thank an anonymous reviewer for their valuable feedback 946on the robustness of our empirical models, particularly regard-947ing potential spatial autocorrelation/dependence in property 948prices, endogeneity, and omitted variable bias.949models4 (SLMs) (Models 18–20) to incorporate a spatial 950lag (SL) term, 951(952)9531954ln955m956j957j958Price959=∑960, in order to account for spatial 961dependence in housing prices (Equation (2)). The spatial 962lag (SL) captures the price effects of neighbouring prop-963erties, j, sold within two months of the subject property’s 964transaction. m is the number of prior transactions. w is 965the spatial weight that governs the structure of the spa-966tial dependence, defined as 1/d, where d represents the 9674 According to Anselin (1988), spatial lag models address omitted 968variable bias by incorporating the spatial dependence between 969observations, where the value of the dependent variable in one 970location may be influenced by values in neighbouring locations. 971By including a spatially lagged dependent variable, the model 972controls for unobserved factors that are shared across proper-973ties, thus reducing the potential for omitted variable bias that 974arises from ignoring spatial correlations (LeSage & Pace, 2009). 975This approach ensures that the effects of local unobserved fac-976tors, which might affect both the dependent variable and inde-977pendent variables, are captured, leading to more accurate and 978unbiased estimates.979Table 7. Socioeconomic ranking models980Model 12981(MDM)982Model 13983(Income)984Model 14985(Employment)986Model 15987(Education)988Model 16989(Access)990Model 17991(Crime)992Constant99311.36294994(0.126073)***99511.30127996(0.128444)***99711.31977998(0.126192)***99911.341821000(0.125968)***100110.818661002(0.132222)***100310.975231004(0.083030)***1005ASB1006–0.0137231007(0.000376)***1008–0.0131021009(0.000393)***1010–0.0128401011(0.000368)***1012–0.0124971013(0.000364)***1014–0.0057181015(0.000502)***1016–0.0109561017(0.000384)***1018ASB × ASB10195.20×10–51020(2.71×10–6)***10215.20×10–51022(2.71×10–6)***10234.49×10–51024(2.72×10–6)***10254.14×10–51026(2.72×10–6)***10274.82×10–51028(2.85×10–6)***10295.57×10–51030(2.86×10–6)***1031Age10320.0015841033(0.000491)***10340.0020531035(0.000500)***10360.0024121037(0.000490)***10380.0019541039(0.000490)***10400.0007851041(0.000122)***10420.0025971043(0.000505)***1044Age × Age1045–3.94×10–61046(3.68×10–6)1047–3.94×10–61048(3.68×10–6)1049–1.05×10–51050(3.68×10–6)***1051–8.68×10–61052(3.67×10–6)***1053–1.42×10–51054(3.81×10–6)***1055–1.04×10–51056(3.80×10–6)***1057Size10580.0079351059(0.000167)***10600.0081681061(0.000170)***10620.0080021063(0.000168)***10640.0077041065(0.000168)***10660.0050291067(6.64×10–5)***10680.0083211069(0.000170)***1070Size × Size1071–7.59×10–61072(3.77×10–7)***1073–8.00×10–61074(3.84×10–7)***1075–7.71×10–61076(3.77×10–7)***1077–7.26×10–61078(3.77×10–7)***1079–8.30×10–61080(3.90×10–7)***1081–8.15×10–61082(3.76×10–7)***1083Parking onsite10840.0375091085(0.010415)***10860.0375091087(0.010415)***10880.0558761089(0.010401)***10900.0369911091(0.010409)***10920.0860691093(0.010858)***10940.0549181095(0.010735)***1096Rural1097–0.0578761098(0.010111)***1099–0.0578761100(0.010111)***1101–0.0673251102(0.010103)***1103–0.0614241104(0.010095)***1105–0.0643121106(0.010782)***1107–0.0824641108(0.010363)***1109Private11100.0302051111(0.023468)11120.0302051113(0.023468)11140.0356401115(0.023485)11160.0422811117(0.023409)***11180.1338551119(0.024722)***11200.1072011121(0.024071)***1122ASB × Z11232.47×10–61124(7.50×10–8)***11252.47×10–61126(7.50×10–8)***11272.24×10–61128(6.94×10–8)***11292.62×10–61130(7.90×10–8)***11314.79×10–71132(1.27×10–8)***11331.24×10–61134(1.03×10–8)***1135Time Dummies1136Yes1137Yes1138Yes1139Yes1140Yes1141Yes1142LGD Dummies1143Yes1144Yes1145Yes1146Yes1147Yes1148Yes1149Property Type 1150Dummies1151Yes1152Yes1153Yes1154Yes1155Yes1156Yes1157EPC Dummies1158Yes1159Yes1160Yes1161Yes1162Yes1163Yes1164Ext Dummies1165Yes1166Yes1167Yes1168Yes1169Yes1170Yes1171Obs1172145211173145211174145211175145211176145211177145211178R211790.67557011800.67557011810.67471311820.67596611830.65133911840.6546871185Adj R211860.67447111870.67447111880.67361111890.67486811900.65015811910.6535661192F1193614.96591194614.96591195612.56831196616.07841197551.70331198583.82371199Prob(F)12000.00000012010.00000012020.00000012030.00000012040.00000012050.0000001206Notes: LGD = Local government district; EPC = Energy Performance Certificate; Ext = Condition of external repair. *** indicates 1% statistical significance. 120712081209===== PAGE 12 (score 48, numeric-paren 23, table-words 8, var-words 2) =====121012114801212D. Lo et al. Property price and anti-social behaviour: An empirical study of Northern Ireland1213Euclidean distance between property i and property j. Ω is 1214a coefficient to be estimated, representing the weighted 1215average effect arising from the spatial dependence of 1216neighbouring properties. Models 18 and 19 are reformu-1217lations of Models 1 and 2, respectively, which account for 1218the spatial dependence of property prices.1219(1220)1221(1222)1223112241122511226112271122811229Ln1230 ln1231 1232 ,1233m1234m1235j1236j1237i1238j1239j1240j1241n1242r1243s1244v1245k1246k1247p1248p1249q q1250u1251u1252k1253p1254q1255u1256Price1257c1258w1259Price1260ASB1261S1262LGD1263PT1264T1265Z1266e1267=1268=1269=1270=1271=1272=1273γ1274+1275=1276+1277+ α ×1278+1279β1280+1281+1282γ1283µ1284+1285µ1286+1287∑1288∑1289∑1290∑1291∑1292∑1293Ω1294 (2)1295We further develop Model 20, which follows a two-1296stage least squares (2SLS) regression approach to address 1297potential endogeneity in the relationship between ASB and 1298property prices. In the first stage, we regress ASB (the en-1299dogenous variable) on an instrumental variable– previous 1300ASB values in 2010 – and a set of control variables above-1301mentioned. The instrumental variable, previous ASB, is as-1302sumed to be correlated with current ASB but not directly 1303with property prices, making it a valid instrument. The pre-1304dicted values of ASB from this first-stage regression, 1305ASB , 1306are used as the instrumented variable in the second-stage 1307regression. In the second stage, we estimate the effect of 1308ASB on property prices by regressing the natural logarithm 1309of property prices on the predicted values of ASB (from 1310the first stage) and other control variables. This two-stage 1311approach allows us to isolate the exogenous variation in 1312ASB, removing any bias that may arise from simultane-1313ity or omitted variable bias in the relationship between 1314ASB and property prices. Table 8 presents the results for 1315the three robustness models. For Models 18 and 19, we 1316observe that the estimates for the main variables, includ-1317ing ASB, ASB × ASB, and other control variables, remain 1318largely consistent with those of the non-spatial models 1319(Models 1 and 2) in terms of both the signs and mag-1320nitudes of the coefficients. Additionally, both the R2 and 1321adjusted R2 values show a moderate improvement with 1322Table 8. Results of robustness models1323Model 181324(SLM 1)1325Model 191326(SLM 2)1327Model 201328(2SLS)1329Constant13309.2348661331(0.098423)***133210.034551333(0.122770)***133411.653431335(1.53242)**1336ASB1337–0.0026111338(0.000755)***1339–0.0041221340(0.000978)***1341ASB × ASB13423.42×10–51343(1.61×10–6)***13441345ASB1346–0.0049341347(0.000723)**134813491350× 1351ASB1352ASB13533.72×10–51354(1.43×10–6)***1355Age13560.0013321357(0.000051)***13580.0023441359(0.000007)***13600.0024231361(0.000126)***1362Age × Age1363–1.11×10–51364(3.01×10–6)***1365–1.24×10–51366(1.72×10–6)***1367Size13680.0041371369(6.23×10–5)***13700.0064781371(0.000154)***13720.0053431373(0.000243)***1374Size × Size1375–5.21×10–61376(2.82×10–7)***1377–2.66×10–61378(1.23×10–7)***1379Parking onsite13800.0775161381(0.019778)**13820.0754561383(0.006855)***13840.0439931385(0.009233)***1386Rural1387–0.0431121388(0.009224)***1389–0.0645211390(0.008455)***1391–0.0682331392(0.008433)***1393Private13940.1632191395(0.022415)***13960.1385411397(0.012894)***13980.1122431399(0.015488)***1400SL Term14011.4180731402(0.057915)***14031.2194521404(0.048232)***14051.2183231406(0.048482)***1407Time Dummies1408Yes1409Yes1410Yes1411LGD Dummies1412Yes1413Yes1414Yes1415Property Type Dummies1416Yes1417Yes1418Yes1419EPC Dummies1420Yes1421Yes1422Yes1423Ext Dummies1424Yes1425Yes1426Yes1427Obs142813,421142913,421143013,4211431R214320.69311214330.71551014340.7033231435Adj R214360.69009414370.71144514380.7002381439F1440651.34611441655.34741442647.52131443Prob(F)14440.00000014450.00000014460.0000001447Notes: ** and *** indicate 5% and 1% statistical significance respectively.144814491450===== PAGE 13 (score 11, numeric-paren 0, table-words 1, var-words 9) =====14511452International Journal of Strategic Property Management, 2025, 29(6), 469–48514534811454the incorporation of the spatial lag terms, which exhibit a 1455positive and statistically significant coefficient, indicating 1456spatial dependence in property prices. Lastly, Model 20 1457presents the results of the 2SLS model. The coefficients 1458on the predicted values of ASB (i.e., 1459ASB) as well as its 1460squared terms exhibit values and statistical significance 1461consistent with those of the non-2SLS models, thereby 1462confirming the robustness of our results.14636. Discussion1464The findings of this study provide robust empirical evi-1465dence that ASB exerts a significant and negative influence 1466on residential property prices in Northern Ireland, after 1467accounting for property-level and temporal attributes. 1468These findings resonate with the established literature, 1469which has long identified a negative relationship between 1470socially undesirable behaviours and property values (e.g., 1471Braakmann, 2012). Through the application of hedonic 1472regression models, the analysis demonstrates that areas 1473with higher incidences of ASB experience notable depre-1474ciation in property values, a trend that largely aligns with 1475previous studies conducted in the U.K. and internationally. 1476Specifically, the results underscore that the impact of ASB 1477on property prices is neither constant nor linear across 1478different spatial and socioeconomic settings of the prop-1479erty market. Instead, it is highly dependent on a variety 1480of property-level and district-specific factors. As Bonakdar 1481and Roos (2023) highlighted, the influence of neighbour 1482quality on property prices varies significantly depend-1483ing on the characteristics of both the property and the 1484neighbourhood. This variability suggests that ASB’s effect 1485is, contrary to common perceptions, multifaceted, influ-1486enced by the characteristics of individual properties and 1487the broader social and economic context of the neigh-1488bourhoods in which they are situated. For instance, while 1489all property types experience depreciation in response to 1490ASB, the magnitude of this effect varies significantly, high-1491lighting the complex nature of ASB’s impact on housing 1492markets. This complexity was also observed by Marsden 1493(2018), who noted that common ASB-related neighbour 1494disputes, such as noise and public hygiene issues, are 1495more likely to reduce property values in urbanised set-1496tings with higher density, a point that is central to our 1497study’s findings.1498One of the pivotal findings of this study is that the ad-1499verse impact of ASB on property values is markedly more 1500pronounced in urban areas, with Belfast emerging as a 1501particularly affected locale. This is in line with the argu-1502ments made by Besley and Mueller (2012), who found that 1503urban environments, particularly those with high popula-1504tion densities and historical socio-political divisions, tend 1505to experience greater negative effects from ASB. Urban 1506environments, defined by higher population densities and 1507pronounced social heterogeneity, exhibit a heightened 1508vulnerability to the pernicious effects of ASB. In Belfast, 1509the city’s diverse demographic fabric–comprising varied 1510religious, cultural, and educational backgrounds–serves 1511as a potential catalyst for ASB. The social tensions and 1512frictions that naturally arise from these differences may 1513exacerbate levels of social disorder, thereby amplifying 1514the prevalence and intensity of ASB. This dynamic under-1515scores the complex interplay between urban diversity and 1516social cohesion, suggesting that the negative externalities 1517of ASB in cities like Belfast are deeply intertwined with the 1518broader sociocultural landscape. Consequently, addressing 1519ASB in such contexts requires a nuanced understanding of 1520the underlying social dynamics that fuel these behaviours.1521The differential impact of ASB on various property 1522types is another noteworthy finding. Apartments, in par-1523ticular, are more adversely affected by ASB in terms of 1524property valuation, whereas detached houses experience 1525the least impact. Several factors may explain this disparity. 1526As Ceccato and Wilhelmsson (2011) pointed out, apart-1527ment buildings are more susceptible to the negative ef-1528fects of social disorders due to shared spaces that fos-1529ter communal disturbances, such as noise nuisance, pets’ 1530barking, and public hygiene concerns. Apartments typically 1531have more communal areas, such as shared hallways, stair-1532cases, and outdoor spaces, which can serve as hotspots for 1533ASB. Moreover, issues like noise nuisance, pets’ barking, 1534and defecation are more common and noticeable in apart-1535ment settings due to the closer proximity of housing units. 1536The lower tolerance levels for such disturbances among 1537apartment dwellers, combined with the higher visibility of 1538ASB in these settings, likely contribute to the more sub-1539stantial price reductions observed in apartment markets. 1540In contrast, detached houses, often located in less dense 1541and more socially stable and prestigious areas, are less ex-1542posed to these issues and therefore less impacted by ASB.1543Privately built houses also display a markedly greater 1544resilience against the price-depressing effects of ASB com-1545pared to their publicly developed counterparts. This could 1546be attributed to better management practices and strat-1547egies typically employed in private developments. Such 1548developments usually have access to more substantial re-1549sources and possess stronger incentives to proactively ad-1550dress and mitigate issues related to ASB. Conversely, public 1551housing estates, frequently situated in socioeconomically 1552disadvantaged areas, are often hindered by insufficient so-1553cial infrastructure and underfunded management, which 1554impairs their ability to effectively counteract the adverse 1555consequences of ASB. This observation underscores the 1556pivotal role that governance and management practices 1557within housing developments play in determining the ex-1558tent to which ASB can influence property values. Therefore, 1559the resilience of privately built homes against ASB-induced 1560price depreciation may be indicative of superior govern-1561ance structures that prioritise the maintenance of property 1562value and community well-being.1563The study further elucidates the subtle interconnection 1564between ASB and the social and economic characteristics 1565of neighbourhoods. This is in line with the literature on 1566socio-economic segregation and residential preferences. 1567As McCann (2013) suggests, high-income households tend 1568to avoid areas with higher levels of ASB, driving a form 1569