Please use this identifier to cite or link to this item: https://hdl.handle.net/10419/322392 
Authors: 
Year of Publication: 
2025
Series/Report no.: 
CBM Working Papers No. WP/4/2025
Publisher: 
Central Bank of Malta, Valletta
Abstract: 
This study examines Malta's peer-to-peer (P2P) rental market by comparing rental listings from Facebook Marketplace and an aggregation of real estate agencies' adverts, while the Housing Authority's official register is used as the benchmark characterising the local rental market. Using a combination of hedonic regression models and natural language processing techniques, including Word2Vec, Doc2Vec, and K-Means clustering, the analysis identifies key differences in pricing strategies, property characteristics, and the role of digital platforms in rental advertising. Findings reveal that advertised listings are skewed towards higherend properties, with a greater prevalence of larger units and higher rental prices compared to the official register. Additionally, real estate agencies exhibit distinct pricing behaviours across platforms, with agency-listed properties on Facebook Marketplace generally featuring advertised rents that are between 1.5% and 2.9% lower than those posted by individual landlords. Moreover, P2P platforms usually contain a broader range of property types compared to those in aggregated agency listings. This suggests that agencies tend to use Facebook Marketplace to reach a wider audience, including lower-budget renters, while agency listings primarily cater to higher-end properties and a wealthier clientele, reflected in its higher average rental prices and greater concentration of premium listings. Linguistic analysis of property descriptions shows that comprehensive, high-quality phrasing increases rental asking prices by 3.7% to 10.2%, while the presence of specific quality-indicative words raises rents by 1.2% to 3.5%, with the range largely depending on the online platform on which an advert is posted. These findings underscore the importance of considering data source biases in rental market analysis and highlight the role of social media and platforms for real estate agencies' listings in shaping rental market dynamics and pricesetting behaviours.
Subjects: 
Rental Market
Peer-to-Peer Market
Advertised Prices
Hedonic Regressions
Natural Language Processing (NLP)
Word2Vec
Doc2Vec
K-Means
Malta
JEL: 
C23
C55
O18
M37
R32
Document Type: 
Working Paper

Files in This Item:
File
Size





Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.