Forecasting models of housing and commercial space demand considering demographic migration trends and mortgage lending availability

Abstract

The study examines approaches to improving the accuracy of demand forecasting for residential and commercial real estate in the largest urban agglomerations of Russia under conditions of intensifying demographic shifts and volatile mortgage lending. The purpose of the study is to develop and empirically test an integrated forecasting model that allows for a quantitative assessment of the impact of migration, household income, and mortgage market parameters on the dynamics of demand and real estate prices. The empirical base is formed from quarterly data series for the period from 2008 to 2023 across fifteen major agglomerations and includes demographic, macroeconomic, and market indicators. Data from Rosstat, the Bank of Russia, relevant ministries, and analytical agencies are used and converted to comparable prices of the base period. The methodology combines correlation and regression analysis, autoregressive and vector autoregressive models with a gradient boosting algorithm implemented in the Python environment using cross-validation procedures on time series. An analysis of pairwise correlations confirms a stable positive relationship between the housing price index and migration balance, as well as a negative dependence on the average mortgage rate, with a moderate influence of real income and construction volumes. Comparative validation shows that the hybrid model based on gradient boosting reduces the mean absolute and root mean square forecasting errors by more than half compared with multiple regression and autoregressive models and provides a determination coefficient of about ninety-five percent on the validation sample. Based on the developed model, demand elasticity with respect to key factors is assessed, showing high sensitivity to the mortgage rate and a more inertial response to migration flows and income changes. Additionally, a three-year scenario forecast for the residential, office, retail, and warehouse segments is conducted, revealing differences in sensitivity to the combination of migration and credit conditions. The study concludes that mortgage affordability plays a dominant role as a short term modulator of demographically driven demand realization, as well as highlights the high practical significance of the developed model for strategic planning by developers, financial institutions, and government authorities in the transition to proactive management of territorial and infrastructure development. The results confirm the high robustness of the proposed approach across different phases of the cycle.