Private consumption, i.e., spending of households, is a key economic variable. While data on private consumption are widely available on a national, aggregate level, disaggregated data on household spending are scarce, particularly in the form of a panel. To fill this gap, we make use of Swiss tax data from the Canton of Bern from 2002 until 2016 to retrieve consumption estimates on a disaggregated level. Since consumption is not directly available from tax records, we show how to transform tax-specific data and information into economically interpretable measures. In particular, we impute consumption based on the simple budget constraint of a household. This approach yields a unique panel of income, wealth and consumption for each taxpayer in the Canton of Bern over time. After discussing and validating the obtained consumption estimates, we analyse consumption and saving patterns of households over the life cycle as well as across different subgroups and show how consumption inequality has evolved over time. We find the typical hump-shaped consumption profile over the life cycle and an increasing savings rate over the working age with a substantial fall with retirement and dissaving thereafter. Our results also suggest that consumption and saving behaviour vary across different household characteristics. Finally, we show that consumption and income inequality remained rather stable between 2002 and 2016. Over the life cycle, however, consumption and income inequality do change: they are rather low at a young age but increase thereafter.
We propose a methodology for measuring the market-implied capital of banks by subtracting from the market value of equity (market capitalization) a credit spread-based correction for the value of shareholders' default option. We show that without such a correction, the estimated impact of a severe market downturn is systematically distorted, underestimating the risk of banks with low market capitalization. We argue that this adjusted measure of capital is the relevant market-implied capital measure for policymakers. We propose an econometric model for the combined simulation of equity prices and CDS spreads, which allows us to introduce this correction in the SRISK framework for measuring systemic risk.