Jonathan Stutz / Fotolia
Jonathan Stutz / Fotolia





Of all the objectives assigned to the welfare state, economic redistribution is a crucial, and sensitive, one. Especially during election times, the topic of taxes and inequality becomes a highly debated subject. Still, if our redistribution preferences as voters are crucial for policy formulation, whence do they arise? Are we really well-informed voters, grounding decisions on truth, or do perception biases also play a role in our political choices? To find out, we teamed up with the Hamburg Institute of International Economics (HWWI) and designed a unique cross-country study. Carried out online as a randomized survey experiment in Brazil, France, Germany, Russia, Spain, Sweden, the United Kingdom, and the United States, it explored citizens’ perception of economic disparities, policy preferences, and responsiveness to new information on the matter.


Our Main Findings in Short:


  • Across all sample countries, a majority of the population experiences difficulties in correctly estimating its position within the income distribution.
  • Except for Brazilians, respondents on average perceive themselves as relatively poorer than they truly are.
  • On top of this, individuals in all countries tend to overestimate (sometimes significantly) the domestic unemployment rate.
  • Participants reporting a lower income see themselves as relatively richer than they actually are, while those reporting a higher income believe they are lower on the income scale than in reality.
  • Our findings suggest that providing people with additional information on income inequality creates a convergence in redistribution preferences across most sample countries.


People tend to think they are poorer than they really are



Figure 1, HWWI 2016


Figure 1 sums up our findings for the average income position bias by country sample. According to our definition, the average person in the samples of Spain, France, the US, and, to a lesser extent, Brazil can be considered as unbiased with values below or close to 10 percentage points. Nonetheless, only Brazil shows on average an overestimation while all other country samples display an underestimation of the income position. The largest negative bias is found for the samples of Germany, followed by the UK, Russia and Sweden. In other words, respondents in all these countries think of themselves as poorer than they truly are.


Yet, this is not the whole story. A mild bias could actually be the result of taking the average across similar numbers of people with a positive and a negative bias. Therefore, Figure 2 shows for each country the distribution of respondents who are either negatively or positively biased, or unbiased.



Figure 2, HWWI 2016


The German sample has the largest share of individuals who underestimate their income position (79%), suggesting that more so than in any other country sample, German respondents tend to place themselves at the lower end of the income distribution. The country samples with high shares without any bias are Sweden, the US, and France (more than 30%). By far the largest share with a positive bias is found for the Brazilian sample (42%). All countries with biases below 10 percentage points in Figure 1 now turn out to have important shares of respondents who misperceive their income distribution


People – sometimes drastically – overestimate unemployment



Figure 3, HWWI 2016


Figure 3 shows that, on average, respondents in all countries tremendously overestimate the unemployment rate, hence, displaying an overly pessimistic view of labour market access. When combining these results with our findings on ranking misperceptions, it appears that people have a harder time assessing the unemployment rate than their own income position. Noteworthy exceptions are the samples of Sweden, which shows a pattern of being the country with the highest share of unbiased individuals, and of Brazil, which is the only country where respondents overestimate both the own position in the income distribution the unemployment rate.


How does your income level influence your perception?


In line with previous findings, the nature and extent of people’s biases varies depending on their actual position within the income distribution (see Table 1). Participants reporting a lower income tend to see themselves as richer than they actually are (positive bias) while those reporting a higher income see themselves as poorer than in reality (negative bias). Individuals in the middle of the income distribution are, in most cases, fairly unbiased.



Table 1, HWWI 2016


The effect of knowing the truth about inequality


Would telling people about the true state of inequality in their country change their judgments on redistribution and the issue of inequality?


Our survey design allowed investigating this question for a reduced country sample (excluding Sweden and the UK). After reporting their income and giving an estimate of their position in the income distribution, half of the participants (treatment group) in each country received information on the state of income inequality in their country and their own true position in the income distribution. The results seem to suggest that respondents who learn that they hold higher income positions than originally assumed tend to favour larger income differences and more personal responsibility. Contrary to that, respondents who are informed about a lower ranking demand smaller income differences and more government interventions.


What can we learn from this?


First and foremost, information helps. When voters’ perceptions of existing inequality are flawed, providing a corrective update on the topic can improve their judgment on redistributive policies. While Germany was the only country where we could notice a significant change across the complete sample in people’s opinions, we find across all country samples that the reaction to the treatment differs by the type of position bias. Thus, reactions to the provision of information depend on the individual context. While estimation biases on income and thus preferences on redistribution differ between upper and lower income percentiles, for all to understand their economic needs and abilities, it is crucial that both top and bottom are adequately informed. This would ensure that the redistribution outcome best matches voters’ preferences.