The Ming and Qing dynasties, according to Wikipedia, citing the Cambridge History of China, collected only 3-4% as a percentage of income for land taxes.
Yet they had extensive bureaucracies going down to the centrally appointed country magistrates and large "permanent" armed forces in the hundreds of thousands (or over a million). How did they pay for it all with only 3-4% tax?
Were things being done without pay or forced extremely low pay? Who would do the jobs then?
Or was 3-4% only on paper and there were other costs borne by the populace for governance?
The short answer is that by the mid-1850s, they couldn't, but then found alternatives.
You must excuse me for failing to give exact figures; I'm on holiday and can't access many of my books, including, irritatingly, those which would be most useful – feel free to prod me in mid-January for that.
Anyway, back to the matter at hand. The Kangxi Emperor would be the last Qing emperor to raise land taxes and managed to maintain solvency, but also, crucially, declared that land taxes would never need to be raised again. His grandson the Qianlong Emperor was, shall we say, unwilling to contradict him, and so the land tax would remain constant for the rest of the Qing Dynasty.^1 ^2 ^3
With no clear source of increase in revenues, it thus became hard to increase expenditure, and so both the army and the civil service were affected. Firstly, the bureaucracy failed to expand to meet population needs. China's population grew from maybe 150 million in 1650 to 300 million in 1800 to 450 million by 1850, but as the number of civil service posts did not increase, by the 1840s some bottom-level magistrates might find themselves with 100,000 people under their jurisdictions. Aside from increasingly inefficient government there was then the issue of reduced social mobility, as young scholars, unable to obtain official positions due to how incredibly competitive they were, became increasingly resentful of the regime and the status quo. The most famous example of this is undoubtedly a Hakka named Hong Huoxiu (better known as Hong Xiuquan), whose exam failures ultimately led to his hallucinations and apparent revelations of messianic status, prompting him to establish the God-Worshipping Society in 1844 and the Taiping Heavenly Kingdom in 1851.^4 But he was not alone. Leading White Lotus practitioners at the time of the White Lotus Revolt of 1796-1806 were in many cases abortive literati, ^5 and among the Taiping leadership at least one more senior member, the Wing King Shi Dakai, is said to have also been a failed examinee (and it is speculated that Feng Yunshan and Hong Rengan may have been similar.)^6 Ultimately I'm not aware of any outright expansion of bureaucratic numbers even later on (which I will get to in a bit).
As regards the army, what happened was in many ways to be expected. Officially the Qing army had around 200,000 Bannermen and 650,000 Green Standard soldiers. In practice many of the Bannermen had slid into comparative irrelevance, particularly in the urban garrisons near the coast, whilst up to three-quarters of Green Banner soldiers officially on the rolls might not even have existed by the time we get to the 1830s.^6 ^7 Mao Haijian's research into Green Banner personnel records indicate that most had a substantial number of dependents and also clearly had other forms of employment, and that, in order to save costs, their actual activity was minimised, such that they were not so much soldiers with trades on the side but tradesmen who did soldiering on the side!^8 An official cost-saving measure was the integration of tax and militia registers. The baojia register served both as a means of collecting financial taxes and of organising militia forces, which would be assembled to deal with domestic threats as they happened rather than serving as a standing force. It is important to note that the main role of the military within Chinese provinces was at this time domestic peacekeeping, and so such systems were able to keep a relative degree of order, at least until the White Lotus and Taiping Rebellions necessitated major reform.^9
And reform they did. Going back to the Wikipedia page, scan your eyes one column to the right, and you will find that little column mentioning commercial taxes. The Qing had three main forms of commercial tax - tariffs, profits from the salt monopoly, and, from 1851 onwards, the likin or lijin, a tax on goods in transit within China. It would be these taxes that bankrolled the Qing for the rest of the dynasty. At the start of the 19th century, land taxes provided around 80% of revenues, commercial taxes the rest. By 1911 it was the exact opposite.^2 What changed? Firstly, the Qing legalised opium in 1858, collecting a 1% tax.^2 Secondly, the introduction of the likin created a huge stream of potential revenue which would only be tapped when the elites collecting it needed (or, as would ultimately be the case, wanted.) The reason was, simply, that the collection of this tax was hugely decentralised, having been implemented in order that provincial generals like Zeng Guofan and Zuo Zongtang would have the means to maintain relative self-sufficiency in their campaigns against the Taiping. In the end, the tax proved useful enough that it was simply never phased out.^9 ^10 Thirdly, arrangements for collecting customs duties were boosted significantly by the establishment of the Imperial Maritime Customs Service, a bureau staffed by Europeans, most notably Sir Robert Hart, who saw themselves as mainly serving Qing interests before those of their home countries.^11 Whilst these measures improved the financial situation it's not difficult to see how central government power was significantly weakened as a result, as authority for half of the country's commercial taxes was placed in the hands of regional and local elites.
How this relates to the Ming is something I've done extensive research on and is still the subject of my current research. The short answer is that under the system envisioned by Hongwu, the state didn't need to pay for most of this.
The Ming had a very small bureaucracy. According to relevant passages in the Ming shilu, in 1371, there were only 1,239 provincial and local officials and 5,488 other personnel. In 1455, the central government seemed to have been staffed by only 1,520 officials. Even with exponential population growth, the total number of civil officials was still only 20,400 persons at the Ming bureaucracy’s fully matured stage. This is an extremely small number of officials given the size of the Ming empire. Hongwu was averse to the notion of a large state bureaucracy dictating the affairs of the common people but would ironically would go on to initiate the largest attempt of state activism in preindustrial China by forcibly relocating tens of thousands of people and organizing the populace into the lijia system. Essentially under the lijia system, the people were supposed to collect taxes, deliver them, and police themselves. The system was designed to immobilize the population and to ensure the self-reproduction of local system. Hongwu envisioned that the lijia would replace the state in local administration.
Sumptuary laws. Under Hongwu's regulations, what the officials could wear, the size of their houses and tombs, and the quality of their eating utensils were contingent upon their rank. Officials were supposed to live entirely on their salaries, which was just enough for sustenance.
Demonetized nature of the economy. Hongwu was an agrarian fundamentalists and his reign marked a return to autarky. The state printed essentially worthless paper currency to defray expenses such as payments to the imperial princes, salaries of the officials, and pay of the soldiers. In theory, Ming salaries were supposed to be paid all with rice but as there was never enough rice to go around, a huge percentage of their pay was converted to paper money instead. Later, when the state used peppers, sappan wood, cloth, and even confiscated materials to pay them. According to the Ming huidian:
For civil and military officials in the capital [Nanjing], [salaries] for rank one and two officials [would be paid with] 40% rice and 60% paper notes; [salaries] for rank three and four officials [would be paid with] 50% rice and 50% paper notes; [salaries] for rank five and six officials [would be paid with] 60% rice and 40% paper notes; [salaries] for rank seven and eight officials [would be paid with] 80% rice and 20% paper notes. Every two ding [one ding in the Ming was equivalent to five guan] of paper notes would be equal to one dan of rice. [Officials of the] relevant departments would go to the respective warehouses by month for disbursement.
The low agricultural tax rate is further misleading because corvee labor was a huge component of the early Ming taxation system. For example, the responsibility for transportation of tax grains to official granaries was borne not by the state but by the people themselves. Similarly, major construction projects such as the construction of Beijing and the repairing of the Grand Canal was done by corvee labor as part of their tax obligations. The state needed to spend very little.
Of course, such a system was unsustainable. In the military, for example, there was large-scale desertion because the military farming colonies could never produce enough food. During the Yongle reign, the Ming had on paper almost 3 million troops but by the early 1430s it was reported that as many as half had deserted. The conversion of official salaries drastically lowered their value. This was because the conversion rate between paper money and material goods were unilaterally set by the state without any regards to the market. The following are from cases I've gathered from the Ming shilu.
This meant that the Ming officials were severely underpaid and Ming officials continually complained about it, especially in the 1470s and 1480s when salaries reached their lowest point.
As a follow-up question, how much of that 3-4% actually went to government coffers? How much would be absorbed by bureaucrats along the way?