Experimental Design
The survey begins by eliciting respondents’ recent motivations for portfolio trades (which can include changes in inflation). We then ask about perceptions of current inflation and expectations about future inflation. The pre-treatment section ends with estimations of unconditional historical asset returns, followed by historical asset returns in periods of increasing and high inflation. We interpret differences in the two return estimations as respondents’ beliefs about assets’ inflation-hedging properties. We elicit return estimates for the German stock market, German energy stocks, the US and Japanese stock markets, two- and ten-year German government bonds, a commodity index, gold, and private real estate.
Following the prior-beliefs section, we randomly allocate each respondent to one of four groups (with equal probability). The first group is a control group that is not shown any information. Group 2 receives information on the time series of inflation, on the possibility of a further increase in inflation, and on drivers of the recent increase in inflation. Group 3 receives actual returns of the German stock market, German energy stocks, the US and Japanese stock market, ten-year German government bonds, and gold in historical periods of increasing and high inflation. This treatment also includes a short summary statement on the past asset returns. Group 4 receives the information provided to group 2 and group 3.
All respondents then answer questions on their one-year- and five-year-ahead expectations about inflation, on hypothetical portfolio choices, on asset-return expectations, as well as on their expectations about risks to the German financial market, their financial situation, and economic development. We also elicit beliefs about consequences of inflation for corporate outcomes and various demographic characteristics. In the months following respondents’ survey completion, we use bank data to test for treatment effects on actual portfolio choices and savings and consumption decisions.
We hypothesize that (i) the inflation treatment increases concerns about inflation, leading respondents to adjust their asset-return expectations and portfolio choices so as to incorporate (prior) beliefs about assets’ inflation-hedging properties; (ii) the asset-return treatment induces expectations and portfolio choices that reflect the newly acquired information on the assets’ past performance during inflationary periods, at least for those concerned about inflation going forward; and (iii) the provision of information about inflation and past returns fosters effects similar to (ii), but more strongly so given increased worries about inflation.