~11 min · pareto, wealth, inequality, power-law, real-world
Level 0Stats Novice
0 XP0/55 lessons0/14 achievements
0/100 XP to next level100 XP to go0% complete
"The wealth distribution is a power law. Every 'average wealth' statistic you have ever read is misleading citizens who do not know to ask for the median or the top-percentile breakdown."
Pareto's Original Observation
The Italian economist Vilfredo Pareto observed in the late 1800s that in Italy, roughly 20% of the population owned roughly 80% of the land. The '80/20 rule' got popularized far beyond its origins, but the underlying mathematical structure — a power-law distribution of wealth — has been confirmed in every modern economy that has been measured. Different societies have different α values for the power law, producing different effective concentration ratios, but the power-law shape itself is universal.
What This Means in Plain Numbers
In modern developed economies, rough orders of magnitude:
The top 1% typically holds 30-45% of total household wealth.
The top 0.1% typically holds 10-25%.
The top 0.01% (the genuine ultra-wealthy) holds a meaningful share that varies more by country.
The bottom 50% typically holds 1-5% of total wealth.
Dad has analyzed the Korean wealth distribution and confirmed what is structurally inevitable in any modern economy: it is power-law-shaped. The 'average wealth' figure that headlines love is dragged enormously by the top end and bears almost no relationship to what the median household actually holds. Korean households at the 50th percentile hold a small fraction of the country's 'average' wealth, just like everywhere else.
Why the Power Law Persists
Three mechanisms keep wealth distributions in their power-law shape:
Preferential attachment: wealth earns returns (interest, dividends, capital appreciation); having more wealth produces proportionally more new wealth.
Inheritance: wealth concentration persists across generations because it transfers across them.
Network effects: high-net-worth individuals access investment opportunities, tax structures, and risk-mitigation tools unavailable to the median household.
The same three mechanisms also explain why the power-law shape is remarkably stable across centuries: it is not a temporary inequality that will 'resolve' on its own. It is the natural attractor of an economy with the dynamics listed above.
The Citizen Lens
Every 'average X for adult Korean / American / etc.' statistic for income, wealth, or assets is silently dragged by the top end of a power-law distribution. The median and the top-percentile distribution are the honest numbers. The mean is a number nobody actually has, except the very few whose presence at the top creates it. The next time a headline cites a national 'average,' translate it: is the underlying distribution bell-shaped (heights, IQ, measurement errors) or power-law-shaped (wealth, income, asset returns, viral reach)? The answer changes the meaning of every claim.
Code
Synthetic Pareto wealth: mean vs median vs top-percentile·python
import numpy as np
rng = np.random.default_rng(220)
# Synthetic 'household wealth' for a country, drawn from a Pareto-like distribution.
N = 1_000_000 # one million synthetic households
alpha = 1.2 # power-law tail exponent
xmin = 10_000 # minimum household wealth
wealth = xmin * (1 + rng.pareto(alpha, size=N))
print(f"Synthetic country: {N:,} households")
print(f"\nMean wealth: {wealth.mean():>15,.0f}")
print(f"Median wealth: {np.median(wealth):>15,.0f}")
print(f"Top 10% threshold: {np.quantile(wealth, 0.90):>13,.0f}")
print(f"Top 1% threshold: {np.quantile(wealth, 0.99):>13,.0f}")
print(f"Top 0.1% threshold: {np.quantile(wealth, 0.999):>12,.0f}")
print()
sorted_wealth = np.sort(wealth)[::-1]
for pct in (0.001, 0.01, 0.10, 0.50, 0.90):
k = max(int(N * pct), 1)
share = sorted_wealth[:k].sum() / wealth.sum()
print(f"Top {pct*100:>5.1f}% hold {share*100:>5.1f}% of total wealth")
# Notice the gap between mean and median. A headline citing the mean as
# 'typical wealth' would be off by an order of magnitude. The honest reporting
# would name the median AND the top-percentile concentration.
Find a recent 'national average X' figure for your country (income, wealth, charitable giving, savings, healthcare spending — any quantity that is plausibly power-law-shaped). Look up the median for the same figure. Compute the ratio mean/median. If the ratio exceeds 1.5, the distribution is meaningfully skewed and 'average' is misleading. If it exceeds 2-3, 'average' is structurally a top-end artifact and the median is the honest number.
Hint
Wealth: mean often 2-5x median in developed economies. Charitable giving: ratio can be 10x or more (a few large donors dominate). Tech salaries: 1.3-1.8x in mid-range, much higher when comp includes equity.
Progress
Progress is local-only — sign in to sync across devices.