Essays on Expectations: Information, Formation and Outcomes

The first chapter investigates the relationship between economic media sentiment and individuals’ expectations and perceptions about economic conditions. We test if economic media sentiment Granger-causes individuals’ expectations and opinions concerning economic conditions, controlling for macroeconomic variables. We develop a measure of economic media sentiment using a supervised machine learning method on a data set of Swedish economic media during the period 1993–2017. We classify the sentiment of 179,846 media items, stemming from 1,071 unique media outlets, and use the number of news items with positive and negative sentiment to construct a time series index of economic media sentiment. Our results show that this
index Granger-causes individuals’ perception of macroeconomic conditions. This indicates that the way the economic media selects and frames macroeconomic news matters for individuals’ aggregate perception of macroeconomic reality.

The second chapter investigates if individuals experiencing different socio-economic environments during their formative years have different expectations about future economic conditions. We analyse differences in expectations across five generations of consumers by testing if they have different levels of confidence. The chapter focuses on all the different generations of the 1900s as defined by Howe and Strauss (1997, 2000). In our econometric model, we use the Millennial Generation as a baseline, as this generation is about to make up the largest fraction of consumers in the economy. Contrary to the theory developed by the literature on generations, such as Howe and Strauss, our results show that confidence increases gradually across generations. We find that the Millennials are more confident than generations born in the first half of the 1900s, but similar in confidence to other generations born in the second half of the 1900s.

The third chapter test whether there is an interaction effect between expectations and policy shocks, that is, whether the effect of monetary policy depends on household’s expectations of the future state of the nationwide economy and their own personal economy. We find that a positive monetary policy shock increases household savings, but the effect is weak when households are more optimistic about their own future household finances and stronger when households are more pessimistic. Households’ expectations of the Swedish economy have no impact on their savings decisions or their response to monetary policy shocks.