There is a broad consensus in international statistical organizations such as Eurostat, the European Central Bank, and the International Monetary Fund that price indices should be constructed using transaction data. However, transaction data often lag behind actual market developments in the housing market for new-built properties as prices are typically set months or years before transactions are finalized. We find that for two large Polish cities (Warsaw and Poznan), house price indices (HPIs) for existing properties lead indices for new builds by, on average, eight quarters. In Poland and other countries with large markets for new-builds, this lag can dramatically distort National HPIs. The lag also affects the European Union’s flagship measure of inflation, the Harmonized Index of Consumer Prices (HICP). This is because the HICP includes owner-occupied housing (on an experimental basis) using exclusively transaction data for new-built properties. We illustrate here that the timeliness issue in the price index for new-built properties disappears when preliminary agreements are used instead of transactions in the compilation of the index.