A variance decomposition analysis for Swiss housing and stock returns

In this study, I use a variance ratio test derived from the Campbell-Shiller return decomposition to test whether there is evidence of a bubble in Swiss housing and stock returns for the period 1980 to 2016. Vector autoregressive models (VAR models) containing macro variables, i.e. real interest rates, real per capita Gross Domestic Product (GDP) growth and a term spread variable, and cash flow data are used for the stock and housing market to model cash flow derived returns of the housing and stock market. From the Campbell-Shiller decomposition the unexpected housing and stock return variance is decomposed into cash flow and return news components. This analysis' findings are that while stock return volatility is driven predominantly by fundamentals like dividend growth, the housing market's unexpected return variance can also be explained mainly by the cash flow (i.e. rent growth) news component, although the variance decomposition is not very robust with respect to the underlying VAR-model. Finally, a variance-ratio test suggested by Engsted and Tanggard (2001) is computed to test whether there were bubbles in the housing and/or stock market.