关于「机场柜台的点差为什么这么宽」,《换汇点差、DCC 陷阱与居留天数》给出的定量结论是什么?
结论:第一章把外币现钞当作库存:年交易量 500,000 的冷门现钞保本半点差是 0.1438,50,000,000 时只有 0.0033,波动率几乎不起作用;在目的地 ATM 用多币种卡一次取 30,000,摩擦率 0.0063,低于本国机场柜台的 0.0450。
数据口径:书中的汇率、费率与行程均为示意模型;全部计算由书中程序生成;本书不是税务、移民、法律或投资建议。 书中以标注框区分已发表的证据、可运行的 Python 计算(16 段程序)、原典与作者的假说。
对照数据表
下表数字由本章 Python 程序原样打印,可重新运行复现。 程序只用 Python 标准库与结论吻合的打印值: 0.0033, 0.1438, 500000, 50000000
| yearly volume | vol 0.05: half-spread | stock days | vol 0.12: half-spread |
| 500,000 | 0.1438 | 122.9 | 0.1477 |
| 2,000,000 | 0.0414 | 64.1 | 0.0429 |
| 10,000,000 | 0.0109 | 29.8 | 0.0114 |
| 50,000,000 | 0.0033 | 13.7 | 0.0035 |
查看完整验算程序
# Chapter 1. p01: Why a booth's spread on foreign banknotes is wide. A
# booth must keep a stock of notes. Each delivery costs K = 2,000 (cash
# courier and insurance); holding notes for a year costs h: interest
# 0.04, insurance and loss 0.01, plus a risk premium g * sigma * sqrt(t)
# for the currency's volatility sigma over the average holding time t.
# With the booth's rent share R = 60,000 a year spread over a yearly
# volume V, the half-spread needed to break even is
# (sqrt(2 K h V) + R) / V (the economic order quantity cost),
# solved for h by fixed-point iteration. No sampling.
from math import sqrt
K, R, BASE, G = 2000.0, 60000.0, 0.05, 1.0
def half_spread(v, sigma):
h = BASE
for _ in range(100):
q = sqrt(2 * K * v / h)
t = q / 2 / v
h = BASE + G * sigma * sqrt(t)
return (sqrt(2 * K * h * v) + R) / v, q / v * 365
if __name__ == "__main__":
print(" yearly volume vol 0.05: half-spread stock days"
" vol 0.12: half-spread")
for v in (500000, 2000000, 10000000, 50000000):
s1, d1 = half_spread(float(v), 0.05)
s2, _ = half_spread(float(v), 0.12)
print(" {:13,d} {:20.4f} {:10.1f} {:20.4f}".format(
v, s1, d1, s2))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《多币种换汇点差微观结构、动态货币转换陷阱与税务居留天数规划》第 1 章「机场柜台的点差为什么这么宽」(版本 1.0.0). 5D Library. DOI 待发布
DOI 待发布 · 专著页正式发布后上线(本站暂无专著页) · 5DL-B-108 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026108,
author = {Fan, Yafu},
title = {Multicurrency FX Spread Microstructure, Dynamic Currency Conversion Traps and Tax-Residency Day-Counting Thresholds},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}