Posted by Bret Devereaux

https://acoup.blog/2025/09/12/collections-life-work-death-and-the-peasant-part-ivc-rent-and-extraction/?utm_source=rss&utm_medium=rss&utm_campaign=collections-life-work-death-and-the-peasant-part-ivc-rent-and-extraction

https://acoup.blog/?p=31305

This is the third piece of the fourth part of our series (I, II, IIIa, IIIb, IVa, IVb) looking at the lives of pre-modern peasant farmers – a majority of all of the humans who have ever lived. Last time, we started looking at the subsistence of peasant agriculture by considering the productivity of our model farming families under basically ideal conditions: relatively good yields and effectively infinite land.

This week we’re going to start peeling back those assumptions in light of the very small farm-sizes and capital availability our pre-modern peasants had. Last week we found that, assuming effectively infinite land and reasonably high yields, our farmers produced enough to maintain their households fairly securely in relative comfort, with enough surplus over even their respectability needs to potentially support a small population of non-farmers. But of course land isn’t infinite and also isn’t free and on top of that, the societies in which our peasant farmers live are often built to extract as much surplus from the peasantry as possible.

But first, if you like what you are reading, please share it and if you really like it, you can support this project on Patreon! While I do teach as the academic equivalent of a tenant farmer, tilling the Big Man’s classes, this project is my little plot of freeheld land which enables me to keep working as a writers and scholar. And if you want updates whenever a new post appears, you can click below for email updates or follow me on Twitter and Bluesky and (less frequently) Mastodon (@bretdevereaux@historians.social) for updates when posts go live and my general musings; I have largely shifted over to Bluesky (I maintain some de minimis presence on Twitter), given that it has become a much better place for historical discussion than Twitter.

Land Holdings

Returning to where we left off last week, we found that our model families could comfortably exceed their subsistence and ‘respectability’ needs with the labor they had assuming they had enough land (and other capital) to employ all of their available farming labor. However, attentive readers will have noticed that the labor of these families could work a lot of land: 30.5 acres for The Smalls, 33.6 acres for The Middles and 56 acres for The Biggs. That may not seem large by the standards of modern commercial farms, but few peasants had anything like such large landholdings; even rich peasants rarely owned so much.

We might compare, for instance, the land allotments of Macedonian and Greek military settlers in the Hellenistic kingdoms (particularly Egypt, where our evidence is good). These settlers were remarkably well compensated, because part of what the Hellenistic kings are trying to do is create a new class of Greco-Macedonian rentier-elites¹ as a new ethnically defined military ruling-class which would support their new monarchies. In Egypt, where we can see most clearly, infantrymen generally received 25 or 30 arourai (17 or 20.4 acres), while cavalrymen, socially higher up still, generally received 100 arourai (68 acres).² That infantry allotment is still anywhere from two thirds to less than half of what our model families can farm and yet was still large enough, as far as we can tell, to enable Ptolemaic Greco-Macedonian soldiers to live as rentier-elites, subsisting primarily if not entirely off of rents and the labor of others.³

Alternately, considering late medieval Europe through the study of Saint-Thibery,⁴ out of 189 households in 1460 in the village just fifteen households are in the same neighborhood of landholdings as the Smalls’ 33.6 acres above (so roughly 55 setérée and up)⁵ only six as much as The Biggs (about 90 setérée and up). In short our assessment so far has assumed our families are extremely rich peasants. But of course they almost certainly are not!

Instead, as we noted in our first part, the average size of peasant landholdings was extremely small. Typical Roman landholdings were around 5-10 iugera (3.12-6.23 acres), in wheat-farming pre-Han northern China roughly 100 mu (4.764 acres), in Ptolemaic Egypt (for the indigenous, non-elite population) probably 5-10 aroura (3.4-6.8 acres) and so on.⁶ In Saint-Thibery in Languedoc, the average (mean) farm size was about 24 setérée (~14.5 acres) but the more useful median farm size was just five setérée (~3 acres); the average is obviously quite distorted by the handful of households with hundreds of setérée of land.

So we might test three different farm sizes; once again, I am going to use Roman units because that’s how I am doing my background math. We might posit a relatively a poor household farm of roughly three iugera (1.85 acres). In Saint-Thibery, 68 of the 189 households (36%) had land holdings this small or smaller, so this is not an unreasonable ‘poor household’ – indeed, we could posit much poorer, but then we’re really just talking about tenant farmers, rather than freeholding peasants. Next, we can posit a moderate household farm of roughly six iugera (3.8 acres); reasonably close to the median holding in Saint-Thibery and roughly what we think of as the lower-bound for ancient citizen-soldier-peasants. Finally, we can posit a large household farm of nine iugera (5.6 acres), reflective of what seems to be the upper-end of typical for those same citizen-soldier-peasants; at Saint-Thibery in 1460 there were a couple dozen families seemingly in this range.⁷

For the sake of a relatively easier calculation, we can assume the same balance of wheat, barley and beans as last time, which lets us just specify an average yield after seed per iugerum of 81.2-189.5 kg of wheat equivalent (achieved by averaging the per-acre wheat equivalent production across all three crops, with seed removed),⁸ with each iugerum demanding between 11 and 15 working days (averaging the labor requirements across all three crops). Finally, we need to remember the fallow: in this case we’re assuming about a third of each farm is not in production in any given year, meaning it is both not consuming any labor nor producing any crops. That lets us then quickly chart out our peasant families based on the land they might actually have (keeping in mind the household size and household land holdings aren’t going to match; the larger household in people won’t always be the one with more land). First, a reminder of the basic labor availability and grain requirements of our households.

Then for the smallest, 3 iugera farm, the numbers work like this:

And then for the medium-sized farm:

And the larger (but not rich peasant) farm:

And we immediately see the problem: only the Smalls manage to get close to subsistence on very favorable (8:1) fertility assumptions on the small farm they own. Now it is possible for the peasants to push a little bit on these numbers. The most obvious way would be focusing as much as possible on wheat cultivation, which has higher labor demands but also the highest yield-per-acre (or iugerum), producing around 50% more calories than beans and 35% more calories than barley per-acre (see last week’s post for specifics). But there’s a limit to going ‘all in’ on wheat to meet food shortfalls: the land might not be suitable for it and wheat exhausts the soil, so our farmers would need some sort of rotation. That said, peasant diets were overwhelmingly grains (wheat and barley) for this reason: they provide the most calories for a favorable balance of land and labor. Our farmers might also try to supplement production with high-labor, high-density horticulture; a kitchen garden can take a lot of work but produce a lot of nutrition in a small space. But hitting household nutrition demands entirely with a kitchen garden isn’t going to work both because of the labor demands but also because the products of a kitchen garden tend not to keep well.

Instead the core problem is that our peasant households are much too large as units of labor for the farmland they own. When we say that, what we mean is that given these households are both units of consumption (they have to provide for their members) and units of production (they are essentially agricultural small businesses), an efficient allocation of them would basically have each household on something like 30 acres of farmland, farming all of it (and thus using most of their labor) and selling the excess. But the lack of economically sustainable social niches – that is, jobs that provide a reliable steady income to enable someone to obtain subsistence – means that these families are very reluctant to leave members without any land at all, so the holdings ‘fractionalize’ down to these tiny units, essentially the smallest units that could conceivably support one family (and sometimes not even that).

I’ve already seen folks in the comments realizing almost immediately why these conditions might make conquest or resettlement into areas of land easily brought under cultivation so attraction: if you could give each household 30-40 acres instead of 3-6, you could realize substantial improvements in quality of life (and the social standing of the farmers in question). And of course that kind of ‘land scarcity’ problem seems to have motivated both ancient and early modern settler-colonialism: if you put farmers next to flat, open ground owned by another community, it won’t be too long before they try to make it farmland (violently expelling the previous owners in the process). This is also, I might add, part of the continual friction in areas where nomads and farmers meet: to a farmer, those grazing fields look like more land and more land is really valuable (though the response to getting new land is often not to create a bunch of freeholding large-farm homesteaders, but rather to replicate the patterns of tenancy and non-free agricultural labor these societies already have to the point of – as in the Americas – forcibly trafficking enormous numbers of enslaved laborers at great cost, suffering and horror, to create a non-free dependent class whose exploitation can enable those patterns. Most conquering armies dream of becoming landlords, not peasants).⁹

Alternately as farms these holdings could be a lot more efficient if they had fewer people on them and indeed when we read, for instance, ancient agricultural writers, they recommend estates with significantly fewer laborers per-unit-land-area than what we’d see in the peasant countryside. But that’s because the Big Man is farming for profit with a large estate that lets him tailor his labor force fairly precisely to his labor needs; the peasants are farming to survive and few people are going to let their brother, mother, or children starve and die in a ditch because it makes their farm modestly more productive per unit labor. Instead, they’re going try to do anything in their power to get enough income to have enough food for their entire family to survive.

There is no real way around it: our peasants need access to more land. And that land is going to come with conditions.

The Big Man’s Land

Now before we march into talking about farming someone else’s land, it is worth exploring why our farmers don’t get more land by just bringing more land under cultivation. And the answer here is pretty simple: in most of the world, preparing truly ‘wild’ land for cultivation takes a lot of labor. In dry areas, that labor often comes in the form of irrigation demands: canals have to be dug out from water sources (mainly rivers) to provide enough moisture for the fields as the most productive crops (like wheat) demand a lot of moisture to grow well. In climates suitable for rainfall agriculture, the problem is instead generally forests: if there’s enough rain to grow grain, there’s enough rain to grow trees and those trees have had quite the head start on you. Clearing large sections of forest by land is a slow, labor-intensive thing and remember, you don’t just need the trees cut down, you need the stumps pulled or burned. Fields also need to be relatively flat – which might demand terracing on hilly terrain – and for the sake of the plow they need to be free of large stones to the depth of the plow (at least a foot or so).

In short, clearing farmland was both slow and expensive and all of this assumes the land can be made suitable and that no one has title to it. Of course if the forest is the hunting preserve of the local elite, they’re going to object quite loudly to your efforts to cut it down. And a lot of land is simply going to be too dry or too hilly or too marshy to be made usable for farming ona practical time-scale for our peasants. Such land simply cannot be brought usefully into cultivation; you can’t farm wheat in a swamp.¹⁰ So it is quite hard and often impractical to bring new land into cultivation.

That doesn’t mean new land wasn’t brought into cultivation, it absolutely was. We can sometimes track population pressures archaeologically by watching this process: forests retreat, new villages pop up, swamps are drained and so on as previously marginal or unfarmable land is brought into cultivation. Note, of course, if you bring a bunch of marginal fields into cultivation – say, a drier hillside not worth farming before – your average yield is going to go down because that land simply isn’t as productive (but demands the same amount of labor!). But that process is generally slow, taking place over generations in response to population pressures. It isn’t a solution available on the time-scale that most of our households are operating. In the moment, the supply of land is mostly fixed for our peasants.

Which means our peasants need access to more land (or another way of generating income). There are a range of places that land could come from:

• Peasant Households without enough labor to farm their own land. In order to make our households relevant at every part of the process, I haven’t modeled the substantial number of very small households we talked about in the first section, households with just 1 or 2 members. If none of those householders were working-age males (e.g. a household with an elderly widow, or a young widow and minor children, etc.) they might seek to have other villagers help farm their land and split the production. For very small households, that might be enough to provide them subsistence (or at least help). Consequently those small, often ‘dying’ households provide a (fairly small) source of land for other households.
• Rich peasants likewise might have more land than their household could farm or cared to farm. Consider the position The Smalls would be if they were a rich peasant household with, say, 25 acres of land (in Saint-Thibery, 26 households (of 189) had this much or more land). That’s enough land that, under good harvest conditions it would be easy enough to shoot past the household’s respectability requirements. At which point why work so hard? Why not sharecrop out a large chunk of your land to small farmers and split the production, so you still make your respectability basket in decent years, but don’t have to work so darn hard?
• The Big Man. Another part of this ecosystem is invariably large landowners, who might have estates of hundreds of acres. Columella , for instance, thinks of farm planning (he is thinking about large estates) in units of 100 iugera (62.3 acres) and 200 iugera (124.6 acres; Col. Rust. 12.7-9). An estate of several hundred acres would hardly be unusual. Likewise in the Middle Ages, the Big Man might be a local noble whose manor estate might likewise control a lot of land. The Big Man might also be a religious establishment: temples (in antiquity) and monasteries and churches (in the Middle Ages) often controlled large amounts of productive farmland worked by serfs or tenants to provide their income. Naturally, the Big Man isn’t doing his own farming; he may have some ‘built in’ labor force (workers in his household, enslaved workers, permanent wage laborers, etc.) but often the Big Man is going to rely substantially on the local peasantry for tenant labor.

In practice, the Big Man is likely to represent the bulk of opportunities here, but by no means all of them. As I noted before, while local conditions vary a lot, you won’t be too far wrong in thinking about landholdings as a basic ‘rule of thirds’ with one third of the land controlled by small peasants, one third by rich peasants and one third by the Big Man (who, again, might be a lord or a big landowner or a church, monastery or temple (in the latter case, the land is owned by the god in most polytheistic faiths) or even the king). But of course only a little bit of the small peasant land is going to be in search of workers, since most peasant households have too many hands for too little land; some of the rich peasant land will be looking for workers (either tenants or hired hands), but rich peasants are still peasants – they do some of their farming on their own. By contrast, the Big Man is marked out by the fact that he doesn’t do his own farming: he needs some kind of labor supply – wage laborers, enslaved/non-free laborers or tenants – for all of it.

But that also means that something like half (or more!) of the land around our peasant village might be owned by a household that needs outside labor to farm it. So we have peasant households with surplus labor that need more land to farm and richer households with surplus land that needs labor. The solution here generally was some form of tenancy which in the pre-modern world generally came in the form, effectively of sharecropping: the landowner agreed to let the poorer household farm some of his land in exchange for a percentage of the crop that resulted. That ‘rent-in-kind’ structure is useful for the peasants who after all are not generally keeping money with which to pay rent. At the same time, it limits their liability: if the harvest on tenant land fails, they may suffer a shortfall, but they aren’t in debt some monetary quantity of rent (though they may end up in debt in some other way).

Now the question is: on what terms?

Tenancy

And the answer here won’t surprise: bad terms. The terms are bad.

There’s a useful discussion of this in L. Foxhall, “The Dependent Tenant” JRS 980 (1990), which in turn leans on K. Finkler, “Agrarian Reform and Economic Development” in Agricultural Decision Making, ed. P.F. Barlett (1980) to get a sense of what the terms for tenant farmers might normally look like. Foxhall notes in this and a few other studies of modern but largely non-industrial farming arrangements that almost no households in these studies were entirely uninvolved in sharecropping or tenancy arrangements, but that the terms of tenancy arrangements varied a lot based on the inputs supplied.

The key inputs were labor, traction (for our pre-industrial peasants, this is “who supplies the plow-team animals”), water and seed. The most common arrangement, representing almost a third of all arrangements, was where the tenant supplied labor only, while traction, water and seed were supplied by the landlord; the tenants share in these arrangements was a measly 18.75%. A number of arrangements had the tenant supplying not only labor but also some mix of traction, water or seed (but not all) and often the tenant’s share of the production hovered between 40 and 60%, with exact 50/50 splits occurring in about a quarter of the sample. In just one case did the tenant supply everything but the land itself; in that case the tenant’s share was 81.25%.

One thing that is obvious from just this example is that arrangements varied a lot and are going to depend on need and bargaining power. A ‘landlord’ who has land they want under cultivation but can supply basically nothing else may be relatively easy to negotiate into a fairly generous deal; a peasant who is absolutely destitute save for the labor of their hands is easy to exploit. An even 50/50 landholder, tenant split seems to have been the norm in much of Europe though, reflected in terms for sharecropper (métayer in French, mezzadro in Italian, mitateri in Sicilian, mediero in Spanish) which all mean ‘halver,’ though again the terms (and the share split) varied, typically based on demand but also on what exactly the landlord was providing (seed, plow teams, tools, physical infrastructure (like a farmhouse), etc).

For the sake of simplicity in our model, we can assume something like a 50/50 split, with our tenants supplying half of the seed, so that our net yield is exactly half of what it would have been. We can then take those assumptions back to our model. To establish a baseline, let’s run the numbers assuming first a ‘medium’ sized (6 iugera, 3.8 acres, with 4 iugera cropped and 2 fallowed) farm, with our fertility estimate set modestly to 6:1, a ‘good but not great’ yield. We’re going to ’round up’ to the nearest even iugerum and assume an average of 13 days per iugerum of labor, just to make our calculations a bit simpler. How hard is it for our peasants to meet their needs if they have to sharecrop the added land they need?

As we can see, tenancy dramatically changes the picture for our peasants. Under these relatively typical assumptions, of our three families all can make subsistence in a normal year but only the Smalls have the right combination of a lot of labor and a relatively small family to have a shot at getting all of their respectability needs (in practice, they’d probably fall short once you consider necessary farm labor not in the fields – fence repair, tool maintenance, home repair and the like). It also isn’t hard to see how we might alter this picture to change our assumptions. Changing the size of the owned farmland has a significant impact (even though it is already so small) because our peasants realize twice the production per unit-land-area for land they own over land they rent (again, terms might vary). Put another way, under these assumptions, somewhat marginal owned farmland that gives an OK-but-not-great yield of 4:1 is of the same use to our peasants as really good tenant-farmed farmland giving a 7:1 yield (both offer 81.2kg of wheat equivalent per iugerum after rent is paid).

That said, the fact that our peasants end up with enough labor to comfortable exceed their subsistence requirements, but not their comfort requirements is favorable for extraction, which we’ll discuss below. These are households with spare labor who can’t fulfill all of their wants entirely on their own, giving the state or local Big Men both a lot of levers to squeeze more labor out of them and also giving the households the available above-subsistence labor to squeeze. By contrast if these peasants had enough land to meet all of their needs themselves, there would be fewer opportunities to compel them to do additional labor beyond that.

But even before we get to extraction, tenancy is also changing our peasants’ incentives. Economics has the concept of diminishing marginal returns, the frequent phenomenon where adding one more unit of a given input produces less and less output per additional input-unit. You will find more errors in the first hour of proofreading than the fiftieth hour, for instance. There’s also the concept of diminishing marginal utility: beyond a certain point, getting more of something is less valuable per unit added. Getting one bar of chocolate when you have none? Fantastic. Getting one bar of chocolate when you have ten thousand? Solidly meh.

Both are working on our farmers to press their natural production inclination not to maximum labor or even hitting that respectability basket but just subsistence and a little bit more. On the diminishing marginal returns front, naturally when it comes to both owned land and rented land, our peasants are going to farm the most productive land first. This is why when we talk about expanding population and expanding agriculture, we often talk about marginal land (less productive land) coming under cultivation; because all of the really great land was already being farmed. But poor farmland doesn’t demand less labor time (indeed, it may demand more), it just produces less. So while we’ve been working here with averages, you should imagine that the first few acres of farmland will be more productive and the latter few less productive.

Tenancy puts this into even more sharp contrast because it creates a really significant discontinuity in the value of farming additional land: the rents are so high that sharecropped or tenant land is much less useful (per unit labor) to the peasant than their own land. So you have a slow downward slope of ‘land quality’ and somewhere in that slope there is the point at which the peasants have farmed all of their own land and so suddenly the effective yield-per-labor-after-rent drops by half (or more!). So the first few hundred kilograms of wheat equivalent are probably fairly easy to get: you have a few good fields you own and your net out of them might be 130-190kg of wheat equivalent per iugerum. Put in a couple dozen days on those two good iugera and The Smalls have just over a quarter of their subsistence needs. But then they have their more marginal fields, which might only yield 80-100kg. Still not terrible but the next couple of dozen days of labor don’t get them as far: not to half but just 44% or so. But now you are out of your own land, so you go to your rich neighbor or the Big Man to get access to some more and suddenly even on their best fields your yield-per-iugerum is 80-95kg so another couple of dozen working days gets you just from 44% to just 57% of what you need. So you need to line up a lot more land, but now you might be starting to look at the worse fields the Big Man has. He still wants them farmed, after all, his choice is between doing nothing and earning money or doing nothing and not earning money; he’d rather earn money. But suddenly you’re looking at maybe as little as 50-60kg of wheat equivalent per iugerum and the labor demands have not gone down.

Meanwhile, the comfort you get from each kilogram of wheat equivalent is also going down. The first 80% or so of your subsistence needs is necessary simply to not starve to death; a bit more makes the household sustainable in the long term. But then – and remember, these choices are coming as you are facing diminishing marginal returns on each day of labor you put in – is it really worth your time to cultivate a couple more fields in order to just get a bit more meat in your diet and have slightly nicer household goods? Wouldn’t you rather rest?

And so what you see is most peasant households aiming not for the full respectability basket, but that “subsistence – and a little bit more” because as each day of labor produces less product and each product produces less joy, at some point you’d rather not work.

And as we’ve seen in theory, our households might hit that crossover point – subsistence and a little bit more – fairly quickly in their labor supply. We haven’t yet, but should now, account for labor spent on things like maintaining tools, fixing fences and other capital investments. If we allocate, say, 45 days, for that and assume that our farmers also want to have some cushion on subsistence (say, another 10%), we might expect The Smalls to be more or less satisfied (on that medium landholding, average 6:1 yields) with something like 245 working days (56% of total), the Middles with 331 working days (65%) and the Biggs with 560 (70%). Working like that, they won’t be rich and won’t ever become rich (but they were never going to become rich regardless), but they’ll mostly survive – some years will be hard – and they’ll have a little bit more time to rest. Some families, a bit more industrious, might push towards achieving most or all of the respectability basket, at least in good years; others might be willing to stick closer to subsistence (or unable to do otherwise).

Of course in areas where the farmland is meaningfully more marginal – average yields around 4:1 rather than 6:1 – our peasants are going to need to work quite a lot more, about 60% more. That pushes the Smalls to about 84% of their available labor, the Middles to 99% and the Biggs actually slightly into deficit, demanding roughly 110% of their available labor. We should keep in mind that each peasant household is going to exist somewhere along the spectrum: some with larger amounts of property or access to better land, some with less. We’ll come back to this in a moment, but this is part of why the poorest of the peasantry were often exempt from things like military service: positioned on marginal land in poor communities, they had little excess labor available. Most peasant households would have been somewhere in between these two, so a labor utilization rate ranging from 50 to 100%, with a lot of households in that 60-80% labor utilization range.

And now you might think, “doesn’t this take us back to peasants actually not working all that much compared to modern workers?” and first I would want to point out that these peasants are also experiencing a quality of living way below workers in modern industrial countries but also no because we haven’t talked about extraction.

Extraction

Because of course the problem here, from the perspective of everyone who isn’t our peasants is that if the peasantry only does the amount of agricultural labor necessary to subsist themselves and just a little more, the society doesn’t have economic room for much else in the way of productive (or unproductive) economic activity. Remember: our peasants are the only significant population actually doing farming. Sure the Big Men and the gentry and temples and monasteries may own land, but they are mostly renting that land out to peasants (or hiring peasants to work it, or enslaving peasants and forcing them to work it).

And those landholding elites, in turn, want to do things. They want to build temples, wage wars, throw fancy parties, employ literate scribes to write works of literature and of course they also want to live in leisure (not farming) while doing this. And the activities they want to do – the temples, wars, fancy parties, scribes and so on – that requires a lot of food and other agricultural goods to sustain the people doing those things. It also requires a bunch of surplus labor – some of that surplus labor are specialists, but a lot of it is effectively ‘unspecialized’ labor.

To do those things, those elites need to draw both agricultural surplus and surplus labor out of the countryside. And we should that of course, obviously, this is an exploitative relationship, but it is also worth noting that for pre-modern agrarian economies, the societies where elites can centralize and control the largest pile of labor and surplus tend to use it to conquer the societies that don’t so ‘demilitarized peasant utopia’ is not a society that is going to last very long (but ‘highly militarized landowner republic’ might).

It is thus necessary to note that when we see the emergence of complex agrarian societies – cities, writing, architectural wonders, artistic achievements and so on – these achievements are mostly elite projects, ‘funded’ (in food and labor, if not in money) out of extraction from the peasantry.

Exactly how this extraction worked varied a lot society to society and even within regions and ethnic and social classes within society. As noted above, in areas where agriculture was not very productive, extraction was limited. By contrast, highly productive regions didn’t so much producer richer peasants as they tended to produce far higher rates of extraction. In some society, where the freeholding farming peasantry (or part of that peasantry) formed an important political constituency (like some Greek poleis or the Roman Republic), the small farmers might manage to preserve relatively more of their surplus for themselves, but often in exchange for significant demands in terms of military and civic participation.

To take perhaps the simplest direct example of removing labor from the countryside, from 218 to 168, the Romans averaged around 10-12 legions deployed in a given year, 45,000-54,000 citizen soldiers.¹¹ Against an adult-male citizen population of perhaps ~250,000 implies that the Roman army was consuming something like a quarter of all of the available citizen manpower in the countryside, though enslaved laborers and males under 17 wouldn’t be captured by this figure. Accounting for those groups we might imagine the Roman dilectus is siphoning off something like 15% of the labor capacity of the countryside on average (sometimes spiking far higher, as much as half of it). On top of that, the demand of these soldiers that they supply their own arms and armor would have pushed farmers to farm a little bit more than subsistence-and-a-little-more to afford the cost of the arms (traded for or purchased with that surplus; at least initially these transactions are not happening in coined money).

We see similar systems in the Carolingian levy system or the Anglo-Saxon fyrd, where households might be brigaded together – in the Carolingian system, households were grouped into mansi – based on agricultural production (you can see how that works above as a proxy for ‘available surplus labor!’) with a certain number – three or four mansi in the Carolingian system – required to furnish one armed man for either a regional levy or the main field army. The goal of such systems is to take the surplus labor above and make it available for military service.

Alternately, the elites might not want their peasants as soldiers but as workers. Thus the very frequent appearance of corvée labor: a requirement of a certain amount of intermittent, unpaid forced labor. This might be labor on the local lord’s estate (a sort of unpaid tenancy arrangement) or labor on public works (walls, castles, roads) or a rotating labor force working in state-owned (or elite-owned) productive enterprises (mines, for instance). As with military service, this sort of labor demand could be shaped to what the local populace would bear and enforced by a military aristocracy against a largely disarmed peasantry. Once again looking at the statistics above, even a few weeks a year per man (rather than per household) would drain most of the surplus labor out of our households. Adding, for instance, a month of corvée labor of per work-capable male (an age often pegged around seven for these societies) under our favorable (6:1) assumptions above bring our work totals to 305 days (70% of total) for the Smalls, 373 (77%) for the Middles and 650 (81.5%) for the Biggs. Corvée labor demands could be less than this, but also often quite a bit more (expectations varied a lot by local laws and customs.

Alternately, elites might just crank up the taxes. In the Hellenistic states (the Ptolemaic and Seleucid kingdoms especially), the army wasn’t a peasant levy, but rather a core of Greco-Macedonian rentier elites (your ‘rich peasants’ or ‘gentlemen farmers’), regional levies and mercenaries. To pay for that (and fund the lavish courts and public works that royal legitimacy required), the indigenous Levantine, Egyptian, Syrian, Mesopotamian (etc. etc.) underclasses were both made to be the tenants on the estates of those rentier elites (land seized from those same peasants in the initial Macedonian conquest or shortly thereafter) but also to pay very high taxes on their own land.¹² So while tax rates on military-settler (that is, Greco-Macedonian rentier elites) land might have been around 10% – 1/12th (8.3%) seems to have been at least somewhat common – taxes on the land of the indigenous laoi could run as high as 50%, even before one got to taxes on markets, customs duties, sales taxes, a head tax and state monopolies on certain natural resources including timber and importantly salt.¹³ So the poor laoi might be paying extortionate taxes on their own lands, lighter taxes on settler (or temple) lands, but then also paying extortionate rents of those tenant-farmed lands.

Another micro-scale option was debt. We’ve been assuming our farmers are operating at steady-state subsistence, but as we keep noting, yields in any given year were highly variable. What peasants were forced to do in bad years, almost invariably as go into debt to the Big Man. But as noted, they’re simply not generating a lot in the way of surplus to ever pay off that debt. That in turn makes the debt itself a tool of control, what we often call debt peonage. Since the Big Man sets the terms of the debt (at a time when the peasant is absolutely desperate) it was trivially easy to construct a debt structure that the peasant could never pay off, giving the Big Man leverage to demand services – labor, tenancy on poor terms, broad social deference, etc. – in perpetuity. And of course, if the Big Man ever wants to expand his land holdings, all he would need to do would be to call in the un-payable debt and – depending on the laws around debt in the society – either seize the peasant’s land in payment or reduce the peasant into debt-slavery.¹⁴

In short, elites had a lot of mechanisms to sop up the excess labor in the countryside and they generally used them.

Consequently, while peasants, unencumbered by taxes, rents, elites, debt, conscription and so on might have been able to survive working only a relatively small fraction of their time (probably around 100 days per year per-working-age male (again, age 7 or so and up) would suffice), they did not live in that world.

Instead, they lived in a world where their own landholdings were extremely small – too small to fully support their households, although their small holdings might still provide a foundation of income for survival. Instead, they had to work on land owned or at least controlled by Big Men: local rentier-elites, the king, temples, monasteries, and so on. Those big institutions which could wield both legal and military force in turn extracted high rents and often demanded additional labor from our peasants, which soaked up much of their available labor, leading to that range of 250-300 working days a year, with 10-12 hour days each, for something on the order of 2,500-3,600 working hours for a farm-laboring peasant annually.

Which is quite a lot less than the c. 250 typical work days (261 weekdays minus holidays/vacation) in the United States – just by way of example of a modern industrial economy – at typically eight hours a day or roughly 2,000 working hours a year. Of course it is also the case that those roughly 2,000 modern hours buy a much better standard of living than what our medieval peasants had access to – consider that a single unimpressive car represents more value just in worked metal (steel) than even many ancient or medieval elites could muster. No, you do not work more than a medieval or ancient peasant: you work somewhat less, in order to obtain far more material comfort. Isn’t industrialization grand?

That said, our picture of labor in peasant households is not complete! Indeed, we have only seen to half of our subsistence basket – you will recall we broke out textiles separately – because we haven’t yet even really introduced the workload of probably the most fully employed people in these households: the women. And what’s where we’ll go in the next post in this series.

https://acoup.blog/2025/09/12/collections-life-work-death-and-the-peasant-part-ivc-rent-and-extraction/?utm_source=rss&utm_medium=rss&utm_campaign=collections-life-work-death-and-the-peasant-part-ivc-rent-and-extraction

https://acoup.blog/?p=31305