There have been two great transformations in human affairs. One is the Neolithic Revolution, the transition from foraging to farming. This is a transformation which is still going on, as there are still some foraging groups around the planet (though it is a vanishing way of life). The second is the Industrial Revolution, the shift from reliance on what is produced by land (farming), a factor of production managed but not created by humans, to reliance on factors of production produced by humans, the produced means of production (capital), such that farmers change from being about 80% of the workforce to less than 5%.
Compare and contrast
Both transformations are technological and involve expanded use of energy, setting off dramatic population increases (the second much faster than the first). In the foraging-to-farming transition, humans no longer merely took food from the environment around them; they deliberately grew food. This food was typically storable, so able to cope with variations in food production across the seasons. Farming both increased the (food) energy to humans and allowed it to be stored for later use, to be actively managed across time. (Hence the very different attitudes to time between foraging and farming cultures.)
Industrialisation used wind, water and (particularly) steam energy to produce things which produced things. Increased agricultural production allowed increased production in general, with expanding sources and use of energy, ushering the creation of, not merely mass prosperity, but increasing mass prosperity. This is in stark contrast for the foraging-farming transition, where it is likely that general standards of living actually fell and, with some exceptions, remained stagnant for thousands of years.
Another contrast is that the foraging-farming transition lead to hierarchical societies with elite-dominated rulerships–whether autocratic, monarchic (i.e. containing powerful noble elites) or deliberative. The last were polities run by elite assemblies, the most democratic of these being some Mediterranean city-states where as much as a third of the adult population got to vote–i.e. male citizens; women, slaves and resident foreigners being excluded. Outside the Mediterranean, assemblies were also important in cities in Lower Mesopotamia and in the kshatriya republics of India. Conversely, with some hiccups, the Industrial Revolution has led to much more broadly-based forms of political life. Another contrast is that the share of output taken by taxation tended to be fairly constant across farming rulerships but has been steadily increasing in modern states.
Why did farming lead to hierarchical societies dominated by controlling elites? The standard answer has been increased production of food led to a surplus above subsistence which allowed a more differentiated society. The problem with this is, why there was any such surplus? Why did not population just increase to consume the surplus? What blocked population increase sufficient to allow the creation of the food surpluses that sustained these elites?
The second problem is, even if there was a food surplus, why did that not just lead to increased specialisation? What happened such that population was blocked from rising to consume the food surplus and that surplus was largely appropriated by a narrow, controlling elite? And, moreover, elites of differing sizes, with different land tenure systems.
Expropriating what you can see
Three Israeli economists have produced a paper (pdf) which provides an elegant answer. Their argument is that the key element is transparency; both in stored food and in expected production. Food had to be stored across the seasons, which made it more vulnerable to expropriation. In their words:
the main aspect of the Neolithic Revolution that gave rise to social hierarchy was not the increase in productivity per se, but rather the induced change in appropriability that accompanied the switch to cereals.
Seen from this perspective, one may acknowledge that not only storage, but also the high population density that resulted in some areas from the transition to agriculture may have made appropriability easier. … it is not that population density created a need for government (as is typically claimed), but rather that higher density may have made the early farmers more vulnerable to exploitation by creating an opportunity for potential exploiters.
The more transparent food production was to some elite, the more it could be appropriated for its purposes. So, in Egypt, where expected production based on simple irrigation techniques was easily visible; where food could be stored with little loss for considerable time; and there was little competition for appropriation of surplus from raiders so little need to create interior fortifications, the ease of taxation enabled a unified, stable autocracy based on landless peasants. The taxman could easily appropriate what could easily be identified–that is, food production was highly transparent to the taxing autocrat (both peasant production and the performance of social intermediaries–nobles, tax officials–in managing and taxing that production). That peasants did not own their land, this was subject to assignment by local headmen controlling the local irrigation, which meant little use of debt (since peasants could not borrow against land) and limited use of law (since local headmen could arbitrate disputes and the rest was executive management). So, even though Egypt was late to farming, it was first to a strong centralised state.
Upper Mesopotamia (later Assyria) came early to farming. Its rain-fed agriculture was both less reliable and less transparent to elites, while raiders were a much greater problem than in Egypt; leading to far more use of peasant land tenure and a waxing and waning of elite control (and population dispersal) due to shifts in relative military efficiency. If farm production is not transparent, the “stick” of dismissal is less effective and the “carrot” of share of the crop more so. If farmers cannot be sacked from the land, it is effectively theirs; they have property rights in land. Rather than working on land allocated to them, they work their own land and pay taxes on it. In the authors’ words:
even when the state has absolute power, an opaque farming technology implies land tenure institutions that are tantamount to farmers owning their land, under the provision that they pay a part of the output to the state. On the other hand, high transparency implies land tenure arrangements that are akin to serfdom, where the cultivator peasant has no title to the land and can be dismissed at the landowner’s will. … [for the farmer] opacity of the production technology provides a shield from the coercive power of the absentee landlord and the state, and thereby provides also some form of freedom.
Under economist Evsey Domar’s well-known analysis of slavery and serfdom, labour bondage is analysed as a product of labour shortage (based on the costs of subsistence plus control being less than free wages so making bondage profitable, free wages being high if labour is scarce compared to land, hence bondage is a product of labour scarcity). While this fits very well the pattern in the Khmer Empire, medieval Europe, early modern Eastern Europe and the colonial Americas, the above analysis suggests, as the authors note, that bondage can arise from different considerations in places of high population density where production is easily observed and so easily expropriated. [Though, strictly speaking, Domar was analysing how and why people become property (slavery), or are bound to property (serfdom), while our authors' analysis is about whether farmers will have property rights in land or be landless. It is also notable that Robert Fogel's analysis of New World slavery stressed the efficiency of the gang system, which closely monitored production according to careful division and allocation of tasks making it much more efficient than free labour in sugar and generally more efficient in rice and cotton production. If unfree production can be monitored and allocated so that the return to the controlling authority is increased over its free equivalent for any practical level of labour density, the Domar analysis does not apply.]
The authors’ wider analysis is very much in line with developments in the economics of information:
Our assumption that the state can dictate the contract to farmers is based on a presumption that the state has already the power to coerce. That power, however, is mitigated here by informational asymmetries. According to our model a transparent production technology implies a ‘stick’ type relation, where the state employs the threat of dismissal to incentivize its agent; under an opaque technology, only a ‘carrot’ type of incentive will be used and the agent is freed from the threat of dismissal.
In Upper Mesopotamia, compared to Egypt, there was also much more use of debt (as land-owning peasants borrowed to tie them over across seasons and regularly found themselves unable to pay back such consumer debt), with the patterns common to land tenure systems of surging debt burdens, peasant loss of land, peasant revolts, declarations of debt-relief. (The authors note the first two features and conjecture land must have flowed back to peasants from poorly managed elite estates; debt amnesties and land redistributions were also a feature of such societies and likely a more important factor.) This led to much more use of law codes to manage the much more complicated social interactions.
Lower Mesopotamia (Sumeria, later Babylonia) was an intermediate case. Farming did not take off until rather more elaborate irrigation than was needed in Egypt or Upper Mesopotamia was established. But city-states arose in Sumeria before Egypt. The elites of local walled cities (as in Upper Mesopotamia, raiders were a problem), including through temple compounds, which were more important than in Upper Mesopotamia, would manage the canal irrigation system, controlling the supply of water to farmers. This made establishing a unified state much more difficult (cities had to be walled due to the raiders and local elites had to manage the canals, so autocracies had much greater difficulties in establishing and maintaining control and suffered greater instability as a result). So, local elites could manage and tax as production was fairly transparent to them, hence the use of landless peasants, but was not so much to any would-be autocrat.
As the authors note, they turn Karl Wittfogel‘s “hydraulic empire” analysis on its head. It is not that irrigation created centralisation in order to manage it, it was that irrigation made peasants more vulnerable to expropriation.
In all these cases, elite expropriation blocked population increase from absorbing all the increased production. The surplus was not what created the elite, the elite created the surplus; probably starting with local leaders managing the protection of fields and food storage from raiders and expanding from there. Such increased need for local leadership to provide protective services that, as a public good, required some expropriation to fund. In Egypt, the original elite role may have come from managing irrigation (pdf). Once you had the kernel of such elites, they then ensured the surpluses continued through appropriation. The simpler the production identification problem (i.e. the more transparent production was to various levels of the elite) the easier it was to tax, to extract the surplus. The more transparent the local process of extraction, the easier for the autocrat to control it. The less contested the extraction by outside raiders, with consequent less need for interior local fortifications, the less potential problems for the autocrat. Hence Egypt had very stable autocratic rule and Mesopotamia not so much.
This process would wax and wane somewhat, but areas would have relatively stable expropriation trade-offs, leading to relatively stable levels of revenues for rulerships. In particular, the more concentrated populations, the cheaper tax collection; the more dispersed populations, the more expensive tax collection. The easier access, the cheaper tax collection–river valleys would be much easier than mountain regions, which would often be not worth the bother.
Close to the opposite end to Egypt was not Mesopotamia, but:
Mediterranean economies that are based entirely on rain and on irregular terrain with idiosyncratic micro-climate (like that of ancient Israel) … with the implication that one should expect there owner occupied farming and weaker states.
The authors’ analysis also helps makes sense of why:
the Pharaohs from early on were considered as incarnations of the gods who regulated the cosmos, while even the mighty kings of Mesopotamia (with a single exception in early Akkad) were only considered as envoys of the gods
The Pharaoh could “see” (and so control) more of social life than the Mesopotamian kings could.
Reading the analysis and considering the social dynamics being dealt with, the order-chaos dichotomy of agrarian religions make a great deal of sense. Chaos was raiders, the rains not coming, the Nile not flooding–things disastrous to be abhorred. Order was peace and plenty (well, not starving). Unsurprisingly, given the reliance on the flooding of the Nile and the transparent social order, the order-chaos dichotomy was particularly strong in Egyptian thinking, with its concept of maat. It was not much of a shift to change that to the good-evil dicohotomy of monotheism.
Contemporary taxation trade-offs
The changes of the Industrial Revolution have increased the capacity of states to observe production and income, hence the rising share of output going to states. To put it another way, what technological changes took place over those millennia did little to change “taxing technology” (i.e. the ability to expropriate). But, as the authors observe, the Industrial Revolution also created more capable states, which did improve the ability to expropriate.
Extending the authors’ analysis further, at what levels of taxation trade-offs between taxing capacity and resistance to same are stable varies from polity to polity. The more mobile taxpayers (or their wealth) are, the greater constraints on taxing. The more say taxpayers have, the easier it is to negotiate higher trade-offs (accepting higher taxes for better services and benefits).
So, highly homogenous societies where there is greater ease of communication between taxpayers and officials, where there is more commonality of preferences, where problems within public provision are more quickly identified and dealt with (improving the efficiency of the tax-benefit trade-off) and where folk may be more reluctant to shift to a “strange” cultural milieu, will have higher levels of taxes. Conversely, more socially diverse societies where there is less commonality of preferences, communication between taxpayers and officials is more fraught, problems within public provision may be less easily identified, agreed upon and (internally) remedied and where movement to a different cultural milieu is less threatening, will tend to have lower levels of taxes. Considerable influx or turnover in population will tend to increase these factors.
Or, in other words, if you want higher government spending, stop immigration and aim for cultural homogeneity. It is notable that Sweden, a highly homogeneous society with notoriously high taxation levels, has become more culturally diverse due to acceptance of (mainly Muslim) migrants. As the above predicts, the result has been a shifting of the tax-benefits trade-off to lower levels of GDP.
Nevertheless, the general tendency seems to be–the more the taxman can see, the more he takes. The more transparent production and income is to the expropriator, the more will be expropriated. It is what constraints there are on taxing which most determines the level of taxation, not some notion of the public good. The latter has far more influence on how taxes are spent, via whose notions of appropriate uses of spending matter, than levels of taxation. (And there is more resistance to taxing than spending, hence the long history of rulerships and states having fiscal problems.) Though there are feedback effects from spending to taxing, in that constraints on taxation include levels of resistance to it and, at the margin, can be affected by what benefits folks are receiving for what cost. Resistance will be lessened the more benefits people that believe they are receiving and the more say they have in what taxes are collected, how and why.
Which is where modern Parliaments arose from; medieval kings needing some forum in which they could make deals to make raising taxes easier. (Getting the merchants to send along elected delegates being a Spanish innovation later picked up by Simon de Montfort and by Edward I.) But expropriation came first; making deals so you could do more of it came second.