Read on Twitter
[1/N] 
New paper! How well do people know their social position relative to others in society? How does their social position shape their views on fairness? Does their income history shape views? With @okoctk & @KBHvidberg Paper here: https://scholar.harvard.edu/files/stantcheva/files/social_position_fairness_nber.pdf
Summary thread
![[1/N] New paper! How well do people know their social position relative to others in society? How does their social position shape their views on fairness? Does their income history shape views? With @okoctk & @KBHvidberg Paper here: https://scholar.harvard.edu/files/stantcheva/files/social_position_fairness_nber.pdfSummary thread](https://pbs.twimg.com/media/Em_URe-WMAA1IsU.png "[1/N] New paper! How well do people know their social position relative to others in society? How does their social position shape their views on fairness? Does their income history shape views? With @okoctk & @KBHvidberg Paper here: https://scholar.harvard.edu/files/stantcheva/files/social_position_fairness_nber.pdfSummary thread")
New paper! How well do people know their social position relative to others in society? How does their social position shape their views on fairness? Does their income history shape views? With @okoctk & @KBHvidberg Paper here: https://scholar.harvard.edu/files/stantcheva/files/social_position_fairness_nber.pdfSummary thread
[2/N] We leverage a unique data combination, linking survey data to detailed administrative data for a large sample of prime-age people in Denmark.
[3/N] The survey gives us perceptions of social position & fairness views; the admin data shows us i) peopleâs true income positions, ii) income histories, iii) past life events, and iv) a lot of info (including income distributions) of many of their possible âreference" groups.
[4/N] We can see where people think they rank & where they actually rank among: people from same cohort, of same sex, living in same municipality, with same education, working in same sector, their immediate neighbors, co-workers in their firm, family members & former schoolmates
[5/N] (Thank you, amazing Danish data! 
)
)
[6/N] We study the relationship between social position & fairness views. And how changes in social position over the course of life & changes due to shocks (unemployment, hospitalization, disability, or promotions) affect views on inequality, fairness, and policy.
[7/N] We also show how a randomized treatment informing individuals of their true positions within reference groups shapes these views.
![[7/N] We also show how a randomized treatment informing individuals of their true positions within reference groups shapes these views.](https://pbs.twimg.com/media/EnBmASYXYAIygEj.png "[7/N] We also show how a randomized treatment informing individuals of their true positions within reference groups shapes these views.")
[8/N] We have three main sets of findings. First, on misperceptions: respondents systematically believe the income levels of others are closer to their own income level than they really are, but, nevertheless, misperceptions are not generally that large.
[9/N] Lower-income respondents tend to underestimate both the median & 95th percentile level of their cohort, while higher-income respondents tend to overestimate these. Same for the income distributions of other reference groups, like people with the same education.
![[9/N] Lower-income respondents tend to underestimate both the median & 95th percentile level of their cohort, while higher-income respondents tend to overestimate these. Same for the income distributions of other reference groups, like people with the same education.](https://pbs.twimg.com/media/EnBl0x3WMAAljYi.png "[9/N] Lower-income respondents tend to underestimate both the median & 95th percentile level of their cohort, while higher-income respondents tend to overestimate these. Same for the income distributions of other reference groups, like people with the same education.")
[10/N] Perceptions of respondentsâ own positions seem biased: lower-income respondents believe they're higher up than they are, higher-income respondents believe they're lower. But caution! This is in part due to mechanical mean-reversion as positions are bounded between 1 & 100.
![[10/N] Perceptions of respondentsâ own positions seem biased: lower-income respondents believe they're higher up than they are, higher-income respondents believe they're lower. But caution! This is in part due to mechanical mean-reversion as positions are bounded between 1 & 100.](https://pbs.twimg.com/media/Em_WhP2WMAMzw6y.png "[10/N] Perceptions of respondentsâ own positions seem biased: lower-income respondents believe they're higher up than they are, higher-income respondents believe they're lower. But caution! This is in part due to mechanical mean-reversion as positions are bounded between 1 & 100.")
[11/N] Re-ranking respondents by perceived position creates an almost perfect match between perceived and actual rank. Even then, more educated respondents are more accurate overall.
![[11/N] Re-ranking respondents by perceived position creates an almost perfect match between perceived and actual rank. Even then, more educated respondents are more accurate overall.](https://pbs.twimg.com/media/Em_WteCW8AADmz4.png "[11/N] Re-ranking respondents by perceived position creates an almost perfect match between perceived and actual rank. Even then, more educated respondents are more accurate overall.")
[12/N] Second, those ranked higher in their cohort think inequality is fairer. This holds for all reference groups: those ranked higher in a group think inequality within group is fairer & that outcomes are result of effort rather than luck. Multiple pieces point to causal link
![[12/N] Second, those ranked higher in their cohort think inequality is fairer. This holds for all reference groups: those ranked higher in a group think inequality within group is fairer & that outcomes are result of effort rather than luck. Multiple pieces point to causal link](https://pbs.twimg.com/media/Em_W9cvXMAA5ULD.png "[12/N] Second, those ranked higher in their cohort think inequality is fairer. This holds for all reference groups: those ranked higher in a group think inequality within group is fairer & that outcomes are result of effort rather than luck. Multiple pieces point to causal link")
[13/N] Respondents' current social position matters more than their past ones for fairness views. But respondentsâ political views are more strongly correlated with their past positions and even with the social position of their parents when they were growing up.
![[13/N] Respondents' current social position matters more than their past ones for fairness views. But respondentsâ political views are more strongly correlated with their past positions and even with the social position of their parents when they were growing up.](https://pbs.twimg.com/media/Em_XF4hWMAAT1J3.png "[13/N] Respondents' current social position matters more than their past ones for fairness views. But respondentsâ political views are more strongly correlated with their past positions and even with the social position of their parents when they were growing up.")
[14/N] Negative income shocks (due to unemployment spells, hospitalization episodes, or disability) decrease peopleâs perceived fairness, while positive income shocks (due to promotions at work) increase it.
[15/N] In our experiment, when we move people's perceived position by correcting their misperceptions, their views on how fair inequality is move as well.
[16/N] Those who overestimated their position and thought they were ranked higher than they truly are switch to thinking that inequality is less fair after being told their true (lower) position; the opposite happens for those who underestimated their social position
[17/N] Third, people view inequality within peers working in the same sector or with same education level as more unfair than inequality among peers of the same age, or same sex, or living in the same municipality.
[18/N] Yet, it is exactly for these reference groups (education & sector) that respondents generally underestimate the degree of inequality most and within which lower-income people strongly overestimate their own positions.
[19/N] Bonus for fun! Respondents perceive their social positions relative to former schoolmates surprisingly well; perceive their positions relative to siblings very well, but far from perfect; & have very little idea about the former social positions of their parents
![[19/N] Bonus for fun! Respondents perceive their social positions relative to former schoolmates surprisingly well; perceive their positions relative to siblings very well, but far from perfect; & have very little idea about the former social positions of their parents](https://pbs.twimg.com/media/Em_X7ckXcAY1UGz.png "[19/N] Bonus for fun! Respondents perceive their social positions relative to former schoolmates surprisingly well; perceive their positions relative to siblings very well, but far from perfect; & have very little idea about the former social positions of their parents")
![[19/N] Bonus for fun! Respondents perceive their social positions relative to former schoolmates surprisingly well; perceive their positions relative to siblings very well, but far from perfect; & have very little idea about the former social positions of their parents](https://pbs.twimg.com/media/Em_X9ftXYAE7H80.png "[19/N] Bonus for fun! Respondents perceive their social positions relative to former schoolmates surprisingly well; perceive their positions relative to siblings very well, but far from perfect; & have very little idea about the former social positions of their parents")
[20/N] Also: our amazing co-author @KBHvidberg is on the job market! Check out his work on his research page http://kristofferhvidberg.com/
