What does Pricing for Lower-Tier Cities conclude, in numbers, about “Lognormal Incomes: Recovering a Whole Distribution from a Mean and a Median”?

Answer: Chapter 2 deals with lognormal incomes and shows that a Pareto tail makes a two-parameter fit overstate the high-income share, 0.0463 against a true 0.0351.

Scope of the numbers: Its city tiers, incomes, price levels, costs and demand parameters are illustrative, and it quotes no specific official statistic; all calculations are generated by its programs; it is not individual business advice. Marked boxes separate published evidence, runnable Python calculations (20 programs), classical sources and the author's hypotheses.

中文版

Data table

The numbers below are printed verbatim by the chapter's Python program and can be reproduced by running it. standard-library Python onlyprinted values that match the answer: 0.0351, 0.0463

alphasigma_fitqsamplelognormalfit/sample
3.00.6050.8065697647940.986
3.00.6050.951003021053491.050
3.00.6050.991538331591231.034
share >= 120,000: sample 0.0259lognormal 0.0314
2.20.6710.8066300682841.030
2.20.6710.951072441170291.091
2.20.6710.991842961848451.003
share >= 120,000: sample 0.0351lognormal 0.0463
Show the full program
# Chapter 2. p22: what a lognormal fit misses when the top is Pareto.
# Illustrative "true" tier: with probability 0.9 a lognormal body
# (median 36,000, sigma 0.55); with probability 0.1 a Pareto tail with
# minimum 60,000 and index ALPHA. 200,000 draws, fixed seed. A lognormal
# is then fitted from the sample mean and median only (as one would from
# a published mean and median), and its upper quantiles are compared
# with the sample's. Parameters are illustrative.
import math
import random
from statistics import NormalDist, median

N01 = NormalDist()
rng = random.Random(20261113)
N = 200000


def draw(alpha):
    out = []
    for _ in range(N):
        if rng.random() < 0.9:
            out.append(36000.0 * math.exp(0.55 * rng.gauss(0.0, 1.0)))
        else:
            out.append(60000.0 * (1.0 - rng.random()) ** (-1.0 / alpha))
    out.sort()
    return out


print("alpha  sigma_fit  q   sample   lognormal  fit/sample")
for alpha in (3.0, 2.2):
    xs = draw(alpha)
    mean = math.fsum(xs) / N
    med = median(xs)
    s = math.sqrt(2.0 * math.log(mean / med))
    for q in (0.80, 0.95, 0.99):
        emp = xs[int(q * N)]
        fit = med * math.exp(s * N01.inv_cdf(q))
        print("%.1f    %.3f   %.2f  %7.0f   %7.0f     %.3f"
              % (alpha, s, q, emp, fit, fit / emp))
    above = len([x for x in xs if x >= 120000]) / N
    fit_above = 1.0 - N01.cdf(math.log(120000 / med) / s)
    print("       share >= 120,000: sample %.4f  lognormal %.4f"
          % (above, fit_above))

Single source and citation

Fan, Y. (2026). Purchasing-Power Parity across Provinces and Pricing for Lower-Tier Cities: Income Distributions, Engel-Coefficient Calibration and Multi-Tier Package Anchors (Version 1.0.0), Chapter 2: Lognormal Incomes: Recovering a Whole Distribution from a Mean and a Median. 5D Library. DOI pending

DOI pending · monograph page goes live after publication · 5DL-B-275 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026275,
  author    = {Fan, Yafu},
  title     = {Purchasing-Power Parity across Provinces and Pricing for Lower-Tier Cities: Income Distributions, Engel-Coefficient Calibration and Multi-Tier Package Anchors},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

Other chapter answers in this book