Adverse selection in iBuyer business models : A study of adverse selection in the use of automated valuation models for iBuyers

Abstract

The purpose of this thesis is to examine how adverse selection can affect the average resale profits for iBuyers, and how simple strategic purchasing rules can help limit this potential problem. The rise of instant buyer (iBuyer) businesses in the past years has made automated valuation models (AVMs) an important part of the property market. Acting as an intermediary between sellers and buyers, the iBuyers provide liquidity and convenience to the market. Although iBuyer services are in demand, large actors within the segment have reported dissatisfying profits over time. In this thesis, hedonic sales, extreme gradient boosting, and support vector machine AVMs are first trained to predict apartment prices in Oslo, Norway. The dataset consists of 84,905 apartment transactions in Oslo, where 80% of the data were used in training. Next, the predictive accuracies of the AVMs are analyzed for different sub-groups of apartments, before purchasing rules are formulated to prevent automated bidding in apartment groups that are hard to price. At last, using the remaining 20% of the data, the average expected resale profits per apartment are examined for a hypothetical iBuyer operating in the Norwegian capital, with and without adverse selection and purchasing rules. We find that adverse selection has a large negative impact on average profits for the hypothetical iBuyer, causing a reduction from 6.29-7.96% to 0.19-1.21% per apartment, across different models and scenarios. Furthermore, the simple purchasing rules are able to limit this reduction with around 1 percentage point per apartment when adverse selection is present. The findings are robust when altering the initial market assumptions, leading to a conclusion that adverse selection poses a noticeable threat to the iBuyer business model. In addition, we conclude that simple purchasing rules can help improve the average profits.