Unlike the large body of research on the determinants of single family prices and rents, the determinants of multifamily rents has received much less exploration. Using a recent and comprehensive micro-level dataset of multifamily housing units in Montgomery County, Maryland, USA, we applied a multilevel linear model with random coefficient to explore the determinants of multifamily rents, including the effects of service and management attributes. The findings are as follows: (1) first we find that a multilevel linear model is better suited to address datasets that include multiple apartment units in a smaller set of facilities, (2) for certain datasets—including ours–a random coefficients model outperforms both an OLS and random intercept model and (3) the effects of service and management variables on multifamily rents vary across types of service and management. Pet allowance, availability of short-term leasing options, and storage service availability increase rents significantly. Conversely, offering units to property employees and services to those with a disability decrease rents significantly.

与有关单身家庭价格和租金的决定因素的大量研究不同，多家庭租金的决定因素受到的探索少得多。我们使用美国马里兰州蒙哥马利县的多户住房单元的最新，全面的微观数据集，应用具有随机系数的多级线性模型来探索多户租金的决定因素，包括服务和管理属性的影响。研究结果如下：（1）首先，我们发现多级线性模型更适合处理包含较少设施中的多个公寓单元的数据集，（2）对于某些数据集（包括我们的数据集），随机系数模型的表现优于OLS和随机截距模型，并且（3）服务和管理变量对多户租金的影响因服务和管理的类型而异。宠物津贴，短期租赁选择的可用性以及仓储服务的可用性显着提高了租金。相反，向物业雇员提供单位和向残障人士提供服务会大大降低租金。