What is it about?
The study investigates why non-local buyers often pay more for homes, focusing on Hong Kong's real estate market from 2010 to 2015. It tests two theories: the Asymmetric Information Hypothesis, where non-local buyers pay more due to higher search costs and lack of local market knowledge, and the Anchoring Bias Hypothesis, suggesting non-locals pay more because prices influence them in their original location. Utilising a novel machine-learning algorithm to identify people's ethnic origin and a large-scale housing transaction dataset, the study employs both a hedonic pricing model and the repeat-sales method for analysis. It finds that non-local buyers generally buy at higher and sell at lower prices than locals, with a shift from a non-local premium to a discount after a new transaction tax, indicating anchoring biases significantly affect pricing decisions.
Featured Image
Photo by Ruslan Bardash on Unsplash
Why is it important?
Our study stands out for its innovative application of advanced AI, specifically a Large Language Model (LLM) like GPT, in analyzing Hong Kong's real estate market from 2010 to 2015. Combining traditional economic theories with modern data analysis techniques is novel in social science research. The use of GPT LLMs for natural language processing and machine learning in parsing a large-scale housing transaction dataset marks a significant methodological advancement. Our findings on the behavioural biases of non-local real estate buyers, mainly focusing on anchoring biases and asymmetric information, offer valuable insights for a diverse audience, including policymakers, economists, and the general public. This research is academically significant and timely, given its alignment with major economic events and policy shifts. By emphasizing the application of GPT LLMs in social science, our study appeals to those interested in the intersection of AI technology and economic behaviour, potentially broadening its impact and readership.
Perspectives
Writing this article was an incredibly enlightening experience for me, as it presented an opportunity to delve into the fascinating intersection of advanced AI technologies, like GPT LLMs, and social science research. What excites me most about this publication is its potential to reshape our understanding of economic behaviours and decision-making processes in the real estate market. Applying GPT-like models back in 2021 before the availability of ChatGPT in parsing complex datasets signifies a leap in methodological approaches and a bridge between technology and social sciences, which often separate academic worlds. This study has reinforced my belief in the transformative power of AI in uncovering nuanced insights into human behaviour, particularly in sectors as dynamic and impactful as real estate. My aspiration is that this research not only contributes to academic discourse but also sparks curiosity and a deeper understanding among a broader audience about how technology can be harnessed to unravel the complexities of economic and social phenomena.
Dr William Cheung
University of Auckland
Read the Original
This page is a summary of: Anchoring and Asymmetric Information in the Real Estate Market: A Machine Learning Approach, Journal of Risk and Financial Management, September 2021, MDPI AG,
DOI: 10.3390/jrfm14090423.
You can read the full text:
Contributors
The following have contributed to this page
