FORECASTING HOUSE RENTAL LEVELS: ANALYTICAL RENT MODEL VERSUS NEURAL NETWORK

The objective of this paper is to investigate the determinants and structure of public sector rents in Hong Kong. It looks at the trends and patterns of public housing rents over the past 20 years or so. The paper discusses the housing authorityʹs setting approach and its iiilierent shurtcommgs with regard to income generation, equity, and effidency. It presents a framework that examines the literature to clarify the underlying factors affecting public housing rents in the context of rent review and rent structure and formulates two interrelated rent models to test hypotheses about these factors. One is a time-series model to ascertain which factors determine rents over time, and the other is a cross-sectional model to quantify the implicit rents of different housing attributes (i.e., the "relativities"). The final section draws out some conclusions about the mechanisms by which the housing authority determines rents and some implications for policy.