The aim of this research project is to investigate how an XGBoost regressor compares to a Random Forest regressor in terms of predictive performance of housing prices with the help of two data sets. The comparison considers training time, inference time and the three evaluation metrics R2, RMSE and MAPE. The data sets are described in detail together with background about the regressor models that are used. The method makes substantial data cleaning of the two data sets, it involves hyperparameter tuning to find optimal parameters and 5foldcrossvalidation in order to achieve good performance estimates. The finding of this research project is that XGBoost performs better on both small and large data sets. While the Random Forest model can achieve similar results as the XGBoost model, it needs a much longer training time, between 2 and 50 times as long, and has a longer inference time, around 40 times as long. This makes it especially superior when used on larger sets of data. / Målet med den här studien är att jämföra och undersöka hur en XGBoost regressor och en Random Forest regressor presterar i att förutsäga huspriser. Detta görs med hjälp av två stycken datauppsättningar. Jämförelsen tar hänsyn till modellernas träningstid, slutledningstid och de tre utvärderingsfaktorerna R2, RMSE and MAPE. Datauppsättningarna beskrivs i detalj tillsammans med en bakgrund om regressionsmodellerna. Metoden innefattar en rengöring av datauppsättningarna, sökande efter optimala hyperparametrar för modellerna och 5delad korsvalidering för att uppnå goda förutsägelser. Resultatet av studien är att XGBoost regressorn presterar bättre på både små och stora datauppsättningar, men att den är överlägsen när det gäller stora datauppsättningar. Medan Random Forest modellen kan uppnå liknande resultat som XGBoost modellen, tar träningstiden mellan 250 gånger så lång tid och modellen får en cirka 40 gånger längre slutledningstid. Detta gör att XGBoost är särskilt överlägsen vid användning av stora datauppsättningar.

Syftet med denna studie var att undersöka hur konsumentvärdet och makroekonomiska faktorer påverkar en bostadsrätts försäljningspris. De variabler som användes för att beskriva konsumentvärdet var våning, månadsavgift, byggår, area och kvadratmeterpris. För att undvika generella prissvängningars inflytande på resultaten används även ett snittpriset inom det undersökta området som ett prisindex. Vidare studeras makroekonomiska faktorers påverkan på prisindexet i en separat analys. I den analysen undersöktes variablerna inflation, BNP-tillväxt, arbetslöshet och reporänta. I analysen av konsumentvärdet visade sig samtliga förklarande variabler ha signifikant relation till försäljningspriset. Analysen för makroekonomiska trender visade enbart ett signifikant förhållande för BNP-tillväxt och inflation i förhållande till prisindexet. Resultaten för prisindexet är osäkra på grund av ett begränsat dataset för de undersökta variablerna samt osäkerhet i reliabiliteten i det datasetet. / The purpose of this study is to investigate how a number of factors connected to the consumer value affects the final sales price of an apartment. More specifically the variables floor, area, cost per area unit and year of construction was examined. To avoid general price fluctuations over time the prices were weighted with a price index for mean prices of apartments in the studied area during the time of sale. Further the price index was studied to find macro economical factors with significant impact on the mean prices. The factors examined in that analysis were inflation, GDP-growth, unemployment rate and official bank-rate. The main method for this study was multiple linear regression. The result yielded by the study showed significant relation between regressors describing consumer value and the final price. The analysis of the macro economical factors showed significant relationship between only inflation and GDP-growth. The result of the regression analysis for the price index were more uncertain due to a smaller dataset and uncertainties in the reliability of that dataset.

Durant les deux dernières décennies, on assiste à l'émergence de Luxemburg comme pôle économique débordant les frontières luxembourgeoise. Cette dynamique se traduit par une croissance migratoire et démographique soutenue entrainant une pression sur le marché résidentiel. On sait à travers la littérature sur la métropolisation qu'elle s'accompagne souvent d'inégalités socio-spatiales mais on en sait encore très peu sur comment ce phénomène se décline dans les espaces transfrontaliers. L’objectif de cette thèse est, dès lors, de comprendre quelles sont les conséquences des marchés fonciers dans l’espace transfrontalier Luxembourg sur la production de l’espace résidentiel ? Le contexte transfrontalier nous incite à aller plus loin pour voir comment les frontières interviennent dans cette production de l’espace. Pour répondre à ces questions, nous avons mené une enquête qualitative auprès de 159 ménages qui a été complétée par une analyse quantitative des flux et des choix de localisation de divers groupes socioprofessionnels en utilisant des modèles de choix discrets. Nos analyses montrent un espace résidentiel luxembourgeois socialement étiqueté avec des lignes de fractures sociales qui s'étendent au-delà des frontières, conséquence du développement économique luxembourgeois et de l'augmentation des prix du foncier. Les prix élevés de l'immobilier constituent une barrière à l'intégration de beaucoup de travailleurs dans le marché résidentiel luxembourgeois. La frontière, par les différentiels qu'elle permet, contribue ainsi à une plus grande spécialisation des territoires et à l'élargissement de l'espace de la métropolisation. Cet élargissement est toutefois limité par les avantages comparatifs qu'offre le Luxembourg à ses résidants. Enfin, malgré l'intégration européenne et un rôle moins prégnant de la frontière en tant que barrière, le processus de métropolisation dans le cas de Luxembourg renforce plutôt le rôle de la frontière en tant qu'instrument de différenciation socio-spatiale. / During the past two decades, Luxembourg has emerged as a small, yet highly specialized metropolis with an expanding cross-border functional area. The economic growth has resulted in increasing migration to Luxembourg and a growing pressure on the housing market. According to the literature, the urban processes in metropolitan regions tend to go hand in hand with socio-spatial inequalities. Previous research has, however, paid less attention to cross-border cities. This observation gives rise to the following question: How does the Luxembourg real estate market contend with its unique cross-border context and what are the socioeconomic effects on the residential space? First, we investigate the comparative advantages of urban land regulations and housing price differentials between the four countries. Second, we conduct a survey on the residential choices on each side of the border. Finally, we analyse cross-border residential mobility of workers using discrete choice models and Luxembourg administrative data. Our analysis shows growing socio-spatial inequalities as a result of the economic development of Luxembourg and the related increase in housing prices. The high housing prices function as a barrier for many to live in Luxembourg. This contributes to a more socially polarized space and an extension of the functional area of Luxembourg beyond the border. However, this extension is somewhat limited and influenced by other residential advantages of living in Luxembourg. Finally, even though the EU integration signifies a declining impact of the border as a barrier, the metropolitanization process in this case reinforces the effect of the border as instrument of socio-spatial differentiation.

published_or_final_version / Housing Management / Master / Master of Housing Management

The thesis is divided into three chapters.1) I study how monetary policy should be optimally designed when households show financial wealth heterogeneity.Main results: thanks to its ability to affect interest payments volatility, monetary policy has real effects even in a flexible-price cashless-limit environment; second, in a setup with nominal rigidities, price stability is no longer optimal. The extent of deviation from price stability depends on the initial level of debt dispersion.2) I assess the role of housing price movements in influencing the optimal design of monetary policy. Under the optimal simple rule, housing price movements should not be a separate target variable in addition to inflation. Furthermore, the welfare loss arising from targeting housing prices becomes quantitatively more significant the higher the degree of access to the credit market.3) I analyze the effects of fiscal policy in a currency area. Results: a public spending shock in one region increases private agents demand for imports and appreciates the terms of trade; second, a countercyclical fiscal rule can restore the Taylor principle, the uniqueness of the equilibrium and reduce macro-volatility.

[Truncated abstract] This thesis examines informational efficiency and price discovery processes within the Perth housing market for the period 1988-2000 by utilising a rich source of Western Australian Valuer General’s Office (VGO) data. Fama’s (1970) classification of market efficiency as potentially weak form, semi-strong, or strong form has been a dominant paradigm in tests of market efficiency in many asset markets. While there are some parallels, the results of tests in this thesis suggest there are also limitations in applying this paradigm to housing markets. The institutional structure of housing markets dictates that a deeper recognition of important housing market characteristics is required. Efficiency in housing markets is desirable in that if prices provide accurate signals for purchase or disposition of real estate assets this will facilitate the correct allocation of scarce financial resources for housing services. The theory of efficient markets suggests that it is desirable for information diffusion processes in a large aggregate housing market to facilitate price corrections. In an efficient housing market, these processes can be observed and will enable housing units to be exchanged with an absence of market failure in all price and location segments. Throughout this thesis there is an emphasis on disaggregation of the Perth housing market both by price and location criteria. Results indicate that the Perth housing market is characterised by varying levels of informational inefficiency in both price and location segments and there are some important pricing-size influences.

This paper uses a factor-augmented vector autoregressive model to examine the impact of monetary policy shocks on housing prices across metropolitan and micropolitan regions. To simultaneously estimate the model parameters and unobserved factors we rely on Bayesian estimation and inference. Policy shocks are identified using high-frequency suprises around policy announcements as an external instrument. Impulse reponse functions reveal differences in regional housing price responses, which in some cases are substantial. The heterogeneity in policy responses is found to be significantly related to local regulatory environments and housing supply elasticities. Moreover, housing prices responses tend to be similar within states and adjacent regions in neighboring states. / Series: Working Papers in Regional Science

A partir de l’évolution des écarts spatiaux des prix du logement et des prix fonciers entre sept secteurs de Bogota, entre 1970 et 2012, cette thèse analyse l’ampleur des recompositions socio-spatiales de la ville. En croisant l’analyse urbaine et l’étude de marchés immobiliers, cette recherche aborde l’évolution des secteurs déjà construits de la ville et les variations des prix des logements anciens. Contrairement aux recherches qui insistent sur la profondeur et la rapidité des transformations des villes latino-américaines sous l’effet de la globalisation, nos résultats montrent que les différences spatiales des prix du logement restent stables. Seuls les écarts spatiaux des prix fonciers augmentent pendant notre période d’étude. L’analyse statistique des facteurs explicatifs de l’inertie des écarts spatiaux des prix montre que la stabilité globale résulte de la conjonction de deux tendances ayant des effets opposés. La diffusion des commerces réduit les écarts spatiaux des prix, alors que la valeur symbolique du secteur de résidence explique le maintien des différences spatiales des prix. Quant aux prix fonciers, la densification des constructions explique une portion des écarts croissants entre secteurs de la ville, mais reste insuffisante pour comprendre l’ampleur des évolutions constatées / Following the evolution of the gaps in housing and land values between seven sectors of Bogotá between 1970 and 2012, this thesis analyzes the scale of urban socio-spatial transformations. By overlapping urban analysis with the study of real estate markets, this analysis deals with the evolution of already-built areas and their housing prices. In contrast with research that emphasizes the depth and rapidity of transformations in Latin American cities during globalization, our findings show that gaps in housing prices from one area to another remain stable. During the period studied, only in the realm of land value gaps in prices between one area and another did actually widen. The statistical analysis of factors explaining the stability of gaps between housing prices from one area to another reveals that overall stability results from the conjunction of two trends with opposite effects. The spatial diffusion of shopping outlets reduces the gaps in housing prices from one area to another, to the same extent as the symbolic or perceived value of each of these residential areas keeps these gaps in prices entrenched. As for land prices, the densification of buildings explains to some extent the widening gaps in prices between some sectors of the city, but this one factor is not enough to explain the extent of the observed evolutions

This study exammes the impact of macroeconomic and financial variables on the performance of the housing property market in South Africa using monthly data for the period January 1996 to June 2008. Orthogonalised and non-orthogonalised house price returns and real estate returns are utilised as proxies for the housing property market in separate models. Three main issues were empirically analysed in relation to the linkage between selected variables and the housing property market. The first aspect examined the relationship between selected macroeconomic and financial factors and property returns. Secondly, the study examined the influence that a unit shock to each variable has on property returns over a period of time. The third aspect focused on determining the proportion of property returns variation that results from changes in the macroeconomic and financial variables. VAR modelling was thus adopted to empirically analyse these three aspects. The results reveal that house price returns are influenced by most of the macroeconomic and financial variables used in this study. Specifically, the real effective exchange rate, interest rate spread and manufacturing production positively impact on house price returns while the domestic interest rate, the dividend yield and expected inflation have a negative effect. Furthermore, manufacturing production has a lagged effect on house price returns while the real effective exchange rate and domestic interest rate have a contemporaneous effect. Real estate returns are not influenced by most of the variables except for the domestic interest rate and dividend yield which have a negative effect.

Cette thèse étudie le marché résidentiel français sous différents angles. Tout d'abord, nous analysons les dynamiques qui régissent les prix des logements en intégrant la dimension démographique. En effet, dans un paysage démographique marqué par l'arrivée des premières générations des baby-boomers à l'âge de la retraite et par le vieillissement plus généralement, il s'avère que la population, que ce soit en terme de taille ou de structure, impacte significativement les prix des logements. L'impact négatif avéré du vieillissement sur les prix est atténué par l'effet positif de la taille de la population. Cependant, la restructuration de la population française en faveurs des personnes âgées n,est pas homogène sur tout le territoire. Dès lors, une approche territorialisée perd tout son sens. En suivant une étude typologique des marchés résidentiels départementaux, le lien démographie-marché du logement apparaît comme facteur structurant dans la différenciation des marchés locaux. Certains départements bénéficieront alors de la création de richesse immobilière, tandis que d'autres subiront une perte amplifiée. En fin, nous avons centré notre analyse sur la métropole du Grand Paris, en considérant l'immobilier résidentiel physique comme une classe d'actif dans un portefeuille de valeurs mobilières. Dans ce contexte, l'actif résidentiel démontre une capacité à protéger contre l'inflation ainsi qu'un pouvoir de diversification notable. Le poids alloué à l'actif résidentiel au sein du portefeuille est significatif quelle que soit la commune ciblée. / This thesis studies the French residential market from different perspectives. First, we analyze the dynamics that govern housing prices by integrating the demographic dimension. Indeed, in a demographic landscape marked by the arrival of the first generations of baby boomers at retirement age and by aging more generally, it turns out that the population, whether in terms of size or structure, has a significant impact on housing prices. The negative impact of aging on prices is moderated by the positive effect iof population size. However, the restructurinf of the French population in favor of the eldery is not homogeneous throughout the territory. Thus, a territorialized study is particularly appropriate. By following a typology approach of departmental residential markets, the demography-housing market link appears as a structuring factor in the differentiation of local markets. Some departments will then benefit from the creation of real estate wealth, while others will suffer an amplified loss. Finally, we focused our analysis on the metropolis of Parisian Region, considering physical residential real estate as an asset class in a portfolio of securities. In this context, residential assets demonstrate an ability to protect against inflation as well as significant diversification power. The weight allocated to residential assets in the portfolio is significant regardless of the targeted commune.