A0343
Title: Valuation of the leasehold properties in the England using a machine learning approach
Authors: Tatiana Franus - Bayes Business School, City, University of London (United Kingdom) [presenting]
Mark Andrew - Bayes Business School City University of London (United Kingdom)
James Culley - Knight Frank (United Kingdom)
Abstract: In England and Wales, the leasehold system divides the ownership of a dwelling into two distinct components for a limited period: leasehold and freehold interests. The leasehold interest pertains to the legal rights of occupation, while the freehold interest encompasses the legal ownership of the land and building. As time passes, the value of the leasehold ownership deteriorates. A novel data-driven approach is introduced to estimate the objective price of the leasehold based on the remaining lease length using a machine learning methodology. Utilizing a dataset of transacted prices from the Land Registry for the period 2010-2016, 183 variables are developed to train machine learning models. The price discounts are estimated for different lease lengths to assess whether they conform to the theoretical prediction that these discounts become steeper as the lease length shortens. From these price discount estimations, the relativity index is then derived, which is the ratio of a finite leasehold value relative to its freehold vacant possession value. The results have the potential to contribute to more equitable outcomes in leasehold extensions, as they enable relativities to be derived from market evidence, providing a more transparent approach to calculating the premium and possessing wide industry implications.
