Impressions on Chen et al (2022)

Title: Banking on the Confucian Clan: Why China Developed Financial Markets so Late Link to the paper.

Authors: Chen, Ma, Sinclair

Year published: 2022

Journal: Economic Journal

This paper is trying to explain why China's banking system developed so late:

It is hard to google "Confucian clans". A better term might be "Chinese kin" or more simply, "clan".
According to Wikipedia, [clans are] "patrilineal and patrilocal group of related Chinese people with a common surname sharing a common ancestor and, in many cases, an ancestral home."
Wikipedia also writes clan as 宗族 or 家族
A Confucian clan is a rigorous clan in that the positions and hierarchy among its members are determined in relation to position in the family and therefore, unchangeable.
In these types of clans, the head of the household (the eldest male member of the family) makes all the decisions in the family and this is unquestioned.
Honor in these clans was obtained by passing difficult imperial exams

The clans can be thought of as corporations but composed entirely of kin
All the financing ever needed by a person who belongs to a clan is obtained within the clan
As such there was no need to develop links outside the clan, including financial links, and therefore (modern) banking between strangers was not necessary.
Even when China opened up [at some year], clans were still dominant
Regions with more clans typically were prosperous and had nice roads and a courier system.

The regression model

The OLS model was \(Banks_i=\alpha+\beta\cdot Clans_i +\gamma X_i +\varepsilon_i\) where

\(i\) is a prefecture index
\(Banks_i\) was bank density in 1936. Bank density is defined as no. of banks per 10000 people.
\(Clans_i\) was geneological density before 1897 (to avoid reverse-causality, they set the year to be so far behind). Geneological density is defined as no. of clan books per 10000 people.
\(X\) is a vector of geographical control variables

In this OLS model, the presence of clans has a negative effect on the bank density
There are other potential explanatory variables they tested which I list here

As usual people complain about omitted variables so this paper uses an instrumental variable

An instrumental variable is a variable that is correlated with the main explanatory variable but not correlated with the other potentially omitted variables
In this case they used each prefecture's shortest distance to then nearest Zhu Xi academy in the 12th century

Zhu Xi was an important dude because he popularized Confucianism amongst the masses from the 12th century onwards
There were only 3 academies that he founded, and the claim is that "his setting them up in those particular locations was because he was "assigned there by fate" one way or another"
Thus the distance of the prefectures to the nearest academy should have nothing to do with the economic and financial development in the early 20th century
Moreover these locations were "neutral" in terms of economic development (when) because these locations were not too good or too bad economically
Also these locations were "neutral" in terms of accessibility (when) because these locations were not too remote (basically, reachable if so desired)

The instrumented \(Clans_i\) turned out to have a greater effect on bank density than the OLS \(Clans_i\).

There are other things they talk about but I skip here for today

Interesting things to look at later

What are the different interest rates among the different clans and was there an idea of equilibrium in this time?