Abstract

It is widely accepted that indexes based on actual transactions are the ideal target measures for most uses, including monetary policy, macroprudential supervision, and financial stability, but also in the national accounts. However, we need to become accustomed to the fact that this first-best solution will, most likely, be unfeasible in practice. This is why valuation-based indicators deserve a closer look, too. Valuation data from investment funds or loan collateralization might be a useful basis for the analysis of prices, rents, and yields (and potentially vacancy rates), particularly in the major cities and prime locations. On the other hand, web-scraped data have proven very useful for residential real estate. In the commercial domain, however, web-scraped data will likely exhibit a (downward) self-selection bias, i.e., only the bottom end of the market would be covered, which is less relevant for macroprudential purposes. A case study compares the distributions of square meterage, prices, rents, yields, and vacancy rates of investment funds data with those of web-scraped data from real estate portals.