基于机器学习方法的一线城市房价影响因素研究

Research on the Factors Affecting Housing Prices in First-tier Cities Based on Machine Learning Methods

一线城市的房地产市场在全国房地产市场中具有举足轻重的地位,因此,维持一线城市房价平稳从而对保障全国房价平稳健康发展具有重要意义。运用XGBoost等机器学习方法和SHAP值可解释性方法,对四大一线城市房价的主要影响因素及其在2012年前后的动态变化进行测算并分析,研究发现:第一,预期因素是一线城市房价上涨的主要影响因素,并且其影响在不断增强。第二,供给因素和需求因素对一线城市房价上涨也起到了较为重要的作用,不过其作用呈现出减弱态势。第三,货币政策等因素对一线城市房价上涨的影响相对偏弱,并且近年来其影响进一步下降。考虑到预期因素是一线城市房价上涨的最主要因素,因此对一线城市而言,稳房价的关键在于稳预期。进一步地,结合实证结果可知,需要让房价更多地由基本面因素来决定,并且通过稳定房地产调控政策来稳预期。一是从供给端发力,构建一线城市土地供给与房价以及土地供给与常住人口之间的联动机制。二是从需求端发力,缩小一线城市与其周边城市以及其他三四线城市之间的公共服务差距,从而减轻一线城市的外来人口压力以及由此引发的住房需求增长。三是保持房地产调控政策的连续性、一致性、稳定性,通过稳定政策来稳定房地产市场的预期。

The real estate market in first-tier cities holds a crucial position in the national real estate market.To promote stable nationwide housing price growth and maintain the stability of housing prices in first-tier cities is of paramount importance.This study combines machine learning methods such as XGBoost with interpretability methods like SHAP values to calculate and analyze the main factors affecting housing prices in the four major first-tier cities and their dynamic changes before and after 2012.Three key conclusions have been drawn:1)Anticipatory factors are the primary drivers of rising housing prices in first-tier cities,and their influence continues to strengthen;2)Supply and demand factors,the two major categories of fundamental factors,also play a significant role in driving up housing prices in first-tier cities,although their impact is diminishing;and 3)Policy factors such as monetary policy have a relatively weak impact on the rising housing prices in first-tier cities,and this impact has further declined in recent years.Considering that anticipatory factors are the most significant drivers of housing price increases in first-tier cities,the key to stabilizing housing prices in these cities lies in stabilizing expectations.Further analysis,in conjunction with the empirical results in this study,suggests the need for housing prices to be determined more by fundamental factors.This can be achieved by stabilizing real estate regulation policies to manage expectations.Three key recommendations are:1)Focus on the supply side by establishing a mechanism linking land supply and housing prices,as well as land supply and the resident population in first-tier cities;2)Address the demand side by reducing the gap in public services between first-tier cities and their surrounding cities and other tier-three and tier-four cities.This can alleviate the pressure on first-tier cities from incoming populations and the resulting increase in housing demand;and 3)Maintain the continuity,consistency,and stability of real estate regulation policies to stabilize expectations in the real estate market.