Affordability Indicators

We measure affordability along two key dimensions of the typical path to homeownership: ongoing mortgage payments and upfront cash requirements. Accordingly, we report two indicators: 

• Mortgage cost ratio (MCR): the average annual mortgage payment relative to the average annual disposable household income 

• Upfront costs ratio (UCR): upfront costs (down payment plus mandatory transaction-related expenses) relative to the average annual disposable household income (see working paper, Sections 3.1 and 3.2). 

Both indicators are ratios relative to annual disposable household income. 

• An MCR of 25 means that an average household spends about 25% of its annual disposable income on mortgage payments. 

• A UCR of 3 means that an average household needs to pay upfront costs amounting to three times its annual disposable income (see working paper, Figure 3). 

The indicators cover the 21 cities included in the German Real Estate Index (GREIX). They are designed to capture urban housing affordability in Germany; rural GREIX counties are not included (see working paper, footnote 6 and Appendix A). 

The sample includes major metropolitan areas as well as smaller urban centres across Germany. Therefore, the indicators are representative for urban housing affordability in Germany. They are not intended to represent affordability in rural areas (see working paper, Section 3). 

We use the average annual disposable household income from the national accounts. Disposable income accounts for taxes and transfers and therefore reflects the amount households can actually spend rather than gross income (see working paper, Section 3.1). 

Regional income data are available for 1995–2022. To obtain a complete time series for 1980–2024, we regress regional income on federal-level income for 1995–2022 and use the estimated relationship to predict regional income for 1980–1994 and 2023–2024. The regression explains more than 96% of the observed variation in regional income (see working paper, footnote 5). 

We re-calculate affordability using alternative measures such as disposable income per capita, model household incomes, and Eurostat income series. While levels can differ, the overall time trends and the key turning points remain unchanged (see working paper, Appendix D.2 and Figures 13–16 and 19–20). 

Aggregate results are computed as weighted averages, using the number of recorded transactions per city and market segment as weights. This implies that larger cities with more transactions receive greater weight. The main time trends remain robust when giving each city equal weight (see working paper, footnote 18 and Appendix D.5, Figures 24 and 25). 

To address potential composition effects from cities joining the sample later, we compute the indicators using a balanced panel of cities observed from 1980 onwards. The results remain very similar to the baseline (see working paper, Appendix D.5 and Figures 26 and 27). 

In the baseline, HP denotes the average house price in the GREIX transaction data for each city, year and market segment (apartments and single-family houses). Using averages reflects the prices of dwellings actually traded in the market (see working paper, Equation (2) and Section 3.1). 

As a robustness check, we construct a hedonic price series using growth rates from the GREIX quality-adjusted indices (anchored at the 2014 average price), which hold important characteristics such as size and housing quality constant. The resulting MCR and UCR series show very similar trends to the baseline, indicating that composition changes do not drive our findings (see working paper, Appendix D.1 and Figures 11 and 12). 

Upfront costs comprise (i) the part of the purchase price not financed by the mortgage (down payment) and (ii) mandatory transaction-related expenses, in particular the real estate transfer tax (Grunderwerbsteuer) and notary fees (including land registry charges). Real estate agent commissions may also add to upfront costs; however, since these fees are not centrally regulated and long-run data are unavailable, they are excluded. Therefore, the UCR should be interpreted as a lower bound of total upfront costs (see working paper, Section 3.2 and footnote 11). 

Because the real estate transfer tax is levied as a percentage of the purchase price, it rises automatically when house prices increase. After the 2006 reform that shifted rate-setting to the federal states, transfer tax rates diverged and rose markedly, making the transfer tax a key driver of total upfront costs in recent years. Notary fees and land registry charges also scale with the purchase price but typically account for a minor share of total upfront expenses (see working paper, Section 3.2). 

Upfront costs increase when borrowers finance a smaller share of the purchase price through a mortgage (lower LTV). However, even under a hypothetical 100% LTV scenario, mandatory taxes and notary fees alone constitute a substantial entry cost for homeownership (see working paper, Section 5 and Appendix E). 

As a robustness check, we compute the MCR assuming a constant LTV of 80% or 100% over time. The resulting series closely matches the baseline in both trend and level (see working paper, Appendix D.3, Appendix E and Figures 21, 22 and 29). 

We perform two robustness checks. First, we compute the MCR assuming a constant term to maturity of 25 years, thereby avoiding the use of repayment rate data. The trend and level remain very close to the baseline (see working paper, Appendix D.4 and Figure 23). Second, we replicate the approach of Biljanovska et al. (2023), obtaining very similar results (see working paper, Appendix D.4 and Figure 28). 

No. Rather than imposing thresholds, we report the mortgage cost ratio and upfront costs ratio relative to disposable household income. This makes it possible to compare affordability over time and across regions in a transparent way (see working paper, Section 1). 

We provide separate indicators for apartments and single-family houses. These segments differ substantially in average prices and financing needs, and they are also the two main homeownership segments covered in the GREIX data (see working paper, Sections 4 and 5). 

Appendix A (Table 1) provides an overview of the years covered for each city, separately for apartments and single-family houses.