ABSTRACT
Some common features are usually obtained with such surveys of inflation expectations and perceptions. Based on the Swedish Household Survey, Jonung (1981) and Palmqvist and Strömberg (2004) report on the demographic properties of inflation perceptions and expectations. They find that low-income households say inflation is higher than respondents in higher income classes, both for perceptions and expectations. Education matters in the same way: the higher the education, the lower the inflation rate reported, and women report higher inflation than men. Furthermore, young and old people report inflation to be higher than the middle-aged (a U-shaped relationship). The same patterns are present in US data. Using the Inflation Psychology Survey, Bryan and Venkatu (2002a) show that reported inflation rates fall with rising income and education, the relationship between age and inflation is U-shaped, and women report higher values than men. In another paper, Bryan and Venkatu (2002b) report similar results for the University of Michigan consumer survey. Increasing income accompanies lower perceived and expected inflation, and women report higher inflation rates than men. One obvious problem with these surveys is that the questions concern variables that are difficult to assess, or even understand. Inflation, for example, is a macro-variable measuring the aggregate price level, but as respondents’ consumption baskets do not necessarily correspond to the one used for calculating the consumer price indices, the answers obtained with the surveys can differ substantially from the official inflation rates. This is often the case; many surveys show perceptions and expectations that differ from the official rates. This chapter investigates [how and whether] incentives might explain the deviations between perceived and expected inflation and the official inflation rate. Since it is costly to gather and analyse data to predict inflation, only those people with strong incentives to do so will attain this information, and thus be able to give informative answers to the surveys. The Joint Harmonised EU Programme of Consumer Surveys has become a unique database, with information on respondents’ views of past and future inflation, but it also contains information on respondents’ likelihood to buy a new car, a house, and making major home improvements, activities that give strong incentives to form correct expectations on inflation. As the dataset used is new, a substantial part of this chapter is devoted to presenting the data, analysing them, and comparing them to other measures of perceived and expected inflation from the same survey, as well as to the results of other surveys. The results will mainly focus on the euro-area aggregate. The next section reviews the dataset. First the answers to the qualitative questions are studied to establish a benchmark for comparison with the quantitative answers. Second, the results from the quantitative answers, broken down by socioeconomic subgroups, are presented. Third, the results are compared with similar breakdowns from other datasets, to see if similar patterns arise in both the national surveys and in the euro-area aggregate. Finally, some preliminary results are given on the time series properties, which are compared with the qualitative data. Section 3 discusses the role of costly information, and how
incentives to collect information matter for respondents when forming their expectations. The empirically obtained answer to the main hypothesis in this chapter, whether incentives to collect costly information improve respondents’ inflation expectation, is given in Section 4. Finally, the last section concludes.
