Purpose
This paper aims to develop a conceptual understanding and a methodological approach for calculating residential net initial yields for both a buy-to-hold and rental investment strategy from hedonic models.
Design/methodology/approach
The markets modelled comprehend of dwellings for rent and sell in Germany. For each of them, two regression models are estimated to extract implicit prices and rents for an artificial identical dwelling and estimate the willingness to pay for the same asset from both a buy-to-hold and rental investment strategy.
Findings
The 3,381 estimated net initial yields in the 161 German markets showed a spatial pattern with the biggest and most attractive cities showing the lowest yields and a self-adjusting process in the markets surrounding the top cities. The net initial yields over time show that prices have increased stronger than rents, leading to rock bottom yields for residential assets and a significant premium in comparison to government bond yields. The approach responds to the spatial hierarchy of markets in Germany, meaning that the level of the estimated yields is accurate and achievable from an investment perspective.
Practical implications
The investment case in residential markets is certainly unique as net initial yields are scarce, especially due to the relatively low number of investment comparables. The paper sheds light on this problem from a conceptual and methodological perspective and confirms that investment yields are deducible by making usage of hedonic models and big data.
Originality/value
In the era of digitalization and big data, residential assets are mostly brought to the market via digital multiple listing systems. Transparency is an essential barrier when assessing the pricing conditions of markets and deriving investment decisions. Although international brokers do provide detailed investment comparables on – mostly commercial – real estate markets, the residential sector remains a puzzle when it comes to investment yields. The paper sheds light on this problem.
