Inequality: winner-take-all markets (superstars) & Income, Inequality and Satisfactions

Implication of power law distributions instead of normal ones

Background: many real-world outcomes (e.g. YT views, city populations, world frequencies) are not normally distributed

→ instead: highly skewed: a few “superstars” or outliers get extremely high values, while most have very low values

• Power law distribution (Zipf’s law, Pareto)

  • ex. small number of yt videos get millions of views; most get very few

  • a few words like “the”, “and”, “of” are used much more than others

Normal distributions underestimate how extreme the “superstar” cases can be

→ For these cases, no expectation of averages or bell curves but extremes

Why do superstars exist?

Theory: demand-side choices and market distribution

• Rational choices: everyone wants the best (most useful word, best video, etc.) → creates a heavy tail where a few caputure most attention (superstars)

  • Ex. most-used words are the most efficient to convey info

  • highly scalable markets (e.g. online videos) reinforce this: sucess breeds success (Zipf distribution power laws)

• Optimization: consumers want high quality which cannot be substituted → winners take most

  • Ex. markets with scalable production (YT, books, etc.) show this most clearly

• Conformity: the more people buy/use something, the more likely new buyers will do the same (“success begets sucess”/matthew effect) & there is a constant low probability of randomly buying new products

  • Ex. Customers sometimes randomly pick new products, allowing for “breakouts”

  • → success of a product as a signaling device for high quality

Study: Bundesliga player market values

Obs: player values are highly skewed, a few stars at the top

Measures:

• market value explained by both performance (goals, assists) and popularity (press citations that are not explained by performance measures)

Results:

• both performance and popularity drive superstar status

• 1 more goal ~4% higher market value

• 1% more press coverage ~0.12% higher value

→ both optimization (quality) and conformity (popularity) matter

Happiness as the ultimate goal

In economics: utility = happines; people strive for well-being

→ well-being typicall measured by wages or leisure, but can also be measured directly

instruments to measure happiness:

• indirectly: via labor market outcomes, health measures (depression, psychological well-being)

• Directly: survey life satisfaction (e.g. SOEP socioeconomic panel, 0-10 scale)

determinants of happiness

positive correlation:

• higher income

• employment

• more education

• having a partner

negative correlation:

• physical or mental illness

→ all results are similar across countries

Income and happiness

individual income is pos. correlated with life satisfaction

But over time: rising income doesn’t always mean more happiness for everyone

Easterlin Paradox:

• At the aggregate level, more income doesn’t guarantee more happiness after a certain threshold

→ Higher income boosts well-being up to a point, then flattens out.

potential explanations for easterlin paradox

social comparison: people care about income relative to others (not just their own)

• Ex. if everyone gets a raise, it’s less special than just you getting a raise

Evidence: social comparison is generally supported, but newer studies question it

→ Problem: hard to define the exact reference group

Potential explanation for easterlin paradox

Adaptation: Pay raises only boost happiness for a short time

→ people adapt (“hedonic treadmill”): new income level becomes the new normal

Empirical evidence:

• 90% adaptation within 3 years

Study: Salary Transparency and Satisfaction

Uni of California: disclosure of public employee wages online & effect on individual satisfaction

Study: Employees randomly informed about the website

Methodology: 20% of control group vs. 50% of treatment group visited the website

Results:

• employees below median pay:

  • lower satisfaction

  • more likely to consider quitting or looking for a new job

• no significant effects for those above the median