Supply decision of existing apartment: Case study of apartment transactions in Gangdong district, Seoul, Korea

Despite its importance, homeowners’ supply type decision has been under-researched due to data limitations. We construct a dataset of 4,037 apartment units in Gangdong district, Seoul, Korea over a 46-quarter period (2011Q1–2022Q2). We simultaneously analyse both the asset and space markets. The results show that the estimated probability and estimated volume of sale from supply type choice models are statistically significant in both the sales and rental price models. The key findings of our study are two. We find that the housing sales and rental markets interact and co-move together. We also demonstrate that the institution of the two-year lease contract period influences the transaction probabilities in both the sales and rental markets. For the supply type decision model, the estimated sale-transaction (rental-transaction) price significantly impacts the sale-transaction (rental-transaction) volume. However, no significant relationship was found between the estimated sale price and rental-transaction volume, nor between the estimated rental price and sale-transaction volume. The results of the forecast-performance comparison show that the integrated supply model_T/probability (calibrated) is the best for sales volume, whereas the two-year lagged value is the best for rental-volume estimation. This demonstrates institutional effects in the rental-housing market in Korea.