We examine how the net worth of billionaires relates to their looks, as rated by 16 people of different gender and ethnicity.
As a group, billionaires are both more educated and better-looking than average for their age. However, when we compare among billionaires, wealth is neither related to beauty nor to educational attainment.
The results do not arise from measurement error or nonrandom sample selectivity. They are consistent with econometric theory about the impact of truncating a sample to include observations only from an extreme tail of the dependent variable.
The point is underscored by comparing estimates of earnings equations using all employees in the American Community Survey to those using a sample of just the top 0.1% of earners.
The findings suggest the powerful role of luck within the extremes of the distributions of economic outcomes. That empirical regularities tend to disappear in the far tails is relevant to analyzing any sample of highly successful or unsuccessful individuals.
[Keywords: financial wealth, education, sample truncation]
…As the bottom panel of Table 1 shows, there was substantial heterogeneity in how the 16 raters perceived the looks of the subjects. The lowest-variance rater included a standardized range of less than two standard deviations, while at the other extreme a rater included a range of over 6 standard deviations.7 Given the idiosyncrasies inherent in the ratings, for each billionaire we first took a simple average of all 16 standardized ratings, with statistics describing this aggregation reported in the third row of the bottom panel of Table 1:8
Because the raters assessed beauty somewhat similarly, the standard deviation of the average of the 16 standardized ratings is below one (as it is in 1998, which used a similar rating scheme). The Cronbach alpha for the 16 raters is 0.88, suggesting substantial agreement among raters, as did the pairwise correlations.
…3. The impact of looks on assets: We relate the billionaires’ assets (in logarithms) to various indexes of their beauty, as assessed by the entire panel of raters. Column 1 of Table 2 presents the simple bivariate relationship between the two. The point estimate is small, non-statistically-significant, and negative—there is no evidence of the positive relationship between economic outcomes and looks that has been so widely demonstrated in the literature.
…The impact of truncating the dependent variable is relevant not only to wealth and earnings. We might expect it to affect studies of the relative success of other elite groups, such as Olympic athletes, Nobel prize winners, Hollywood stars, and Fortune 500 CEOs. At the opposite tail of these distributions, the same statistical difficulty may also affect analyses of the determination of economic outcomes among especially disadvantaged groups, such as the long-term homeless, the persistently unemployed, habitual drug users, and people with chronic health problems. Luck matters more at the extremes. When we sample only from the tails, the world becomes less predictable.