However, at the same time, we do not observe that tax rates (or even military targets) were fine-tuned to respond to local economic shocks, like changes in cash crop prices or a severe drought in 1938.

Accordingly, the colonial authorities exercised caution and moderation in the setting of head tax rates: more affluent districts (closer to seaports, producing cash crops and having a railway line) were set higher rates, while the military target was set only in proportion to local population.

For head tax levy, we show that spikes in head tax rates statistically significantly increased the likelihood of tax-related conflicts, yet were not associated with conflicts of other motives.

Resistance to conscription mainly boiled down to individual defiance through absenteeism at the drafting boards, as manifested by relatively high absenteeism rates across colonies.

Resistance to conscription mainly boiled down to individual defiance through absenteeism at the drafting boards, as manifested by relatively high absenteeism rates across colonies.

Likewise, we find strikingly high compliance with the head tax across all colonies. Even under the most conservative demographic assumptions, the collected head tax revenue consistently exceeded 80% of the theoretical expectation, based on head tax rates and estimates of the eligible population.

Our findings reveal that colonial coercion was highly effective. First of all, military recruitment targets were nearly always met, with the mean Military Enforcement Rate standing at 99.4%!

To do so, we undertook extensive digitization efforts and created a unique district-level panel dataset covering military conscription and head tax levies for every year from 1919 to 1949, for all eight colonies in French West Africa (example of conscription table below).

We contribute to this debate by looking at two pillars of colonial rule in French West Africa: head tax collection and military conscription.

Such status quo playing (choice reluctance) is more pronounced among the Chinese respondents from working-class or farming families, and significantly less visible among individuals with family backgrounds in the private sector. The same heterogeneity doesn’t exist for the French sample.

In regression terms, the significant merit-luck gap starts to appear for the Chinese sample when we either exclude only the status-quo distributions, or all status-quo players from the sample.

But! When we remove the status quo players from the sample, Chinese respondents who do not play status quo (i.e., those who make an active choice) display a significant merit-luck gap in their redistribution decisions. The non-status-quo-playing Chinese are as meritocratic as the French!

Furthermore, Chinese respondents do not redistribute more to the worker who lost out in the luck scenario than in the merit scenario, across both status quo treatments, confirming previous findings that they do appear unmeritocratic in the aggregate, while the French are consistently meritocratic.

We indeed confirm our hypothesis: Compared to French respondents (colored in blue), Chinese respondents (colored in red) consistently and significantly choose more non-redistribution (dotted line) across both highly unequal (12/0 split) and relatively equal status quo (7/5 split) treatments.

We also augment the luck scenarios commonly used in third-party spectator games by introducing additional dimensions of inequality of opportunities, such as head starts, obstacles, Type I and Type II errors, as well as marginal winners.

We run a third-party spectator game with elite university students in China and France, by varying the initial split of payoffs between two real-life workers to redistribute from, i.e., either the respondent is assigned to an equal status quo (7/5), or unequal status quo (12/0) scenario.