Taxation and income distribution

I've been using ACKS! to run some games set in the Hyborean age. I've been thinking of the assumptions of the domain revenue system and applying it to different cultures. What I don't have is some idea on the level of taxation (in percentage) and the distribution of wealth and income. These are some of the stuff hidden in Alex's great spreadsheets. So the questions are: What is the average tax percentage? Are there some fixed taxes too (taxation per head or hearth)? What is the income distribution? Does it match the distribution of levels (and the upkeep required at those levels)? What are the proportions of 0-level income classes (peasants, laborers, craftsmen, professionals, petty aristocracy)?

Hello Esa, those are tough questions. Fortunately, several of these questions have been answered in my recent Patreon articles, which explore the wealth distribution between the government and the landowner, and between lord and peasant, as well as the productivity of the peasantry itself. If you don't like Patreon type services, I summarize everything quarterly, and the "economics and domains" issue of Axioms is coming out in about 30 days.

 

Thanks! I started supporting you on Patreon and will check out your notes on peasant economics. Incidentally I ran some numbers last week based on the Apostolides et. al article from 2008. Unfortunately I don't have access to their primary database, as it would be interesting to see the temporal and spatial variability of the values. I'll compare both on a spreadsheet and see how they look.

OK, here are some numbers for comparing the ACKS! article to two different studies of English agriculture ("English Agricultural Output and Labor Productivity, 1250-1850: Some Preliminary Estimates", Apostolides et al. 2008 and "English and Welsh Agriculture, 1300-1850: Output, Inputs, and Income", Allen 2005), one bottom up approach based on manorial records and the other top-down based on national accounts. I used the data for the 1300-1350 era and will use the values from Apostolides for England unless there are significant differences to Allen

Some values showing significant difference in English data to the ACKS! assumptions:

  • Family size 4.5 people
  • Average farm size 14 acres
  • Pasture per average farm 30-35 acres (including fallow fields)
  • Yields for different crops differ, but are at least 10 bushels per acre (this value probably hides large variability in crop yields for different counties with different crop mixes)
  • Gross seed ratio is at least 5 except for oats
  • The monetary value of crops is 3.9-4.7 gp/acre
  • With larger amount of pasture per farm, the amount of animals per acre are almost 10 times lower
    • The percentage of animals in the pastures are:
    • Horses 1-2%
    • Oxen 2-3%
    • Cattle 6-10%
    • Sheep 81-85%
    • Swine 4-5%
    • NB! No numbers for goats or poultry in either article
    • NB! In the amount of pasture used per animal nothing is overlapping and there's no exceptions for pannage
  • There are 17-38 acres of field per ox, this compares to the often quoted value of 30 acres per ox (one ox and one plow per two neighboring farms)
  • The animal yields per animal are lower in most categories
    • Also the yields/acre are only 10-20% of ACKS! values both because of lower per animal yields and more pasture required per animal
  • The estimates of hide yield per animal is doubled for cattle and 2/3 for sheep
    • Still the per acre yield for cattle is pretty close to ACKS!
    • For sheep it's less than 5%
  • The total monetary value for animal products is 1.1-1.2 gp/acre
  • The total monetary yield per acre is 5-6 gp and for the average farm 70-80 gp (150-180 gp for a 30 acre farm)
  • The energy production per acre is 1300-1600 calories (of which 80-100 are from animals)
    • The assumption for average consumption is lower, only 2000 per person
    • Using that value an acre of farm with around 2 acres of pasture will support 0,6-0,8 people
    • A single farmer will support about the same amount of people as in ACKS and an average sized farm can support 9-11 people (before expenses)
  • The assumptions for labor are very different
    • Annual labor by the family head is 270-330 days, 310-380 for the whole family (Only 13.6% for the spouse and nothing for children and other dependents)
    • This leaves only 10-27% of free days in a year
    • The amount of man-days of work are a lot higher for anything except harvest (6 days/acre)
      • 12 days/acre for arable land
      • 20 days/acre for pastoral land
      • 94 days per livestock unit (= one cow)
    • Calculating from these the maximum size of farm that a single family can work is 5.5 acres (17 acres including pasture)
      • Also a single family can only harvest 5 acres of fields in a month
    • Both values should increase when the rest of the family is taken into account, both for harvest and animal keeping
  • The spread of expenses is different too:
    • Around 50% for labor (feeding the family and hiring help for harvest etc.)
    • 10-11% for capital (half for implements and half for animals)
      • An average farm has only roughly 10 gp worth of animals and 10 gp worth of tools (one set of laborer's tools?)
      • That farm has 0.2 livestock units per acre, 4-6 per farm (that should be around 50 gp so something is wrong)
    • 23-27% for rent
    • 15% for tithes and taxes
  • With these values the average farm can support 4-6 people at an average energy consumption of 2000 calories (Allen gives the actual average energy consumption of 1800 calories). Both could be increased by taking into account poultry, fish, gardens and gathering from forests etc.
    • The average family can run a farm of up to 15 acres of fields (+ pastures) which is pretty close to the average size.
    • That average family has 35-100 days free for other labor (making a maximum of 10 gp), which could also be used for running a larger field or herd.
  • With these English values the total income for farming is only slightly more profitable than working as cheap labor, but also the income for the domain lord is lower (70-90 gp gross annually, roughly 3 gp per month as domain income and 3-4 gp net income for the farmer).
    • Also renting the land at the usual 1/33 per month is too expensive (compared to the historical it's 21 times too much and even at historical levels around half of the domain income)

Thanks for providing the comparison.

 didn't use either Allen or Apostolides as sources. Allen is extensively quoted in one source I did use, British Economic Growth, 1270-1970 (Broadberry, Campbell, et. al, 2010), but they came to very different conclusions. Another main source was Standards of Living in the MIddle Ages (Dyer). I also consulted a lot of sources on Rome, as ACKS is very much inspired by Antiquity. 

Comments below!

  • Family size 4.5 people
    • ACKS is 5 people
  • Average farm size 14 acres; Pasture per average farm 30-35 acres (including fallow fields)
    • What year is being used for this average? British farm versus pasture changes radically changed over time due to the Black Death and other factors
    • ACKS is 30 acres including pasture and fallow fields
    • Keep in mind ACKS is agnostic about whether the family owns the farm - it could be a 15 acre family farm and 15 acres of demesne land, etc.
  • Yields for different crops differ, but are at least 10 bushels per acre (this value probably hides large variability in crop yields for different counties with different crop mixes)
    • Yields vary widely in every source I've seen. ACKS' values are typical of Late Antiquity and the Dark Ages, not as good as the best of the High Middle Ages nor as good as the best in Classical Antiquity.
  • ​The monetary value of crops is 3.9-4.7 gp/acre
    • ACKS is 4gp/acre
    • What did you use as your conversion rate to gp?
  • With larger amount of pasture per farm, the amount of animals per acre are almost 10 times lower
    • ​ACKS' pasture values were sourced from multiple books/articles that had broad agreement.
    • ACKS' pasture values are for mixed-use farming where the animals and crops are working in tandem, with different animals grazing in same pasture on different crops. Transhumance and nomadic pastoralism will have many fewer animals per acre.
  • The assumption for average consumption is lower, only 2000 per person
    • Consumption of 2000 calories per day seems low for an active male laborer; Dyer cites much higher values, and Roman army fed soldiers 3,500 calories per day
  • The assumptions for labor are very different!
    • ​Yes they are! I cross-checked my values across a few different sources, though they do often vary by an order of magnitude.
    • (For what it's worth, Harn Manor also used 500 working days per household)
  • An average farm has only roughly 10 gp worth of animals and 10 gp worth of tools (one set of laborer's tools?)
    • That seems very low
    • What did you use as your conversion rate to calculate gp?
    • Dyer has an average of 1.7 horses/oxes, 2.9 cattle, 11.2 sheep, and 1.85 pigs per household
    • Broadberry has an average of 3.36 cattle, 42 sheep, 2.4 pigs per 30 acres
  •  Renting the land at the usual 1/33 per month is too expensive
    • ​Richard Cantillon indicates rents are typically between 1/3, 2/5, and 3/4 of the total production of the farm depending on the country. England had particularly low rents compared to other times and places.

In any case, the nice thing about ACKS (at this point) is that all the assumptions are clear so you can change them to suit your particular time, place, productivity, etc. for your campaign.

 

 

Hi!

I used these data as they were easily available. I originally got them to make some estimates on development of productivity (for Traveller). There should be similar data for Dutch and Swedish farming, but I havent found any nice overviews yet. I would like to extend these both ways to antiquity and modern time periods. And the English situation was different, especially after the Black death when the labor market changed and the wool wars begun. Comparison to other datasets (especially non-european) would be beneficial. I wonder if there are any studies of historical Chinese agricultural productivity in English.

Some notes on your comments:

  • Average farm size 14 acres; Pasture per average farm 30-35 acres (including fallow fields)
    • What year is being used for this average? British farm versus pasture changes radically changed over time due to the Black Death and other factors
    • These data are based on the 1300-1350 period. Also they are raw averages and don't really tell that much about very small and large farms (both of which were plentiful)
    • ACKS is 30 acres including pasture and fallow fields
    • Keep in mind ACKS is agnostic about whether the family owns the farm - it could be a 15 acre family farm and 15 acres of demesne land, etc.
    • The farm size for zero profit (subsistence farming) varies highly depending on the soil productivity, ease of plowing and climate
      • Some idea can be gathered from traditional "acre" or the amount of land that could be plowed in a day in different areas (from 900 to 12 000 square meters per day)
  • Yields for different crops differ, but are at least 10 bushels per acre (this value probably hides large variability in crop yields for different counties with different crop mixes)
    • Yields vary widely in every source I've seen. ACKS' values are typical of Late Antiquity and the Dark Ages, not as good as the best of the High Middle Ages nor as good as the best in Classical Antiquity.
    • Also the climate should affect these values. Comparing Egyptian crop yield to Scandinavian there's a huge difference. We need more data to know the variability and then we can see how it fits to the 1-3 gp/month for vegetable domain revenue range in ACKS!
  • ​The monetary value of crops is 3.9-4.7 gp/acre
    • ACKS is 4gp/acre
    • What did you use as your conversion rate to gp?
    • I just valued those for the cost of products in ACKS!. Converting actual prices is possible, but not easy. Some authors use conversion to daily labor, some for daily energy consumption and some to amount of silver. The last one sounds most D&D-suitable.
  • With larger amount of pasture per farm, the amount of animals per acre are almost 10 times lower
    • ​ACKS' pasture values were sourced from multiple books/articles that had broad agreement.
    • ACKS' pasture values are for mixed-use farming where the animals and crops are working in tandem, with different animals grazing in same pasture on different crops. Transhumance and nomadic pastoralism will have many fewer animals per acre.
    • Those values were based on the total of all fallow fields, pasture, meadows and commons. Using just fallow fields gives values that are pretty close to the ones in ACKS!. Different assumptions on actual land use. Probably reality is more varied with some animals grazing only on fallow and some using common lands less efficiently. Also both sources assume that most of oats harvested (a third of the crops) were fed to animals. Maybe the more labor intensive pastoralism was more concentrated and the less labor intensive had the animals roaming on commons and were not grazing them as efficiently.
  • The assumption for average consumption is lower, only 2000 per person
    • Consumption of 2000 calories per day seems low for an active male laborer; Dyer cites much higher values, and Roman army fed soldiers 3,500 calories per day
    • The sources I used had the average at 2000, minimum subsistence at 1500 and active adult farm laborer at 4000 calories. With the real energy production per person at 1800 calories in the 1300 era and 3400 in the 1500 era I would say that the farmers in 1300 were famished and probably some of the productivity was lost because of this.
  • The assumptions for labor are very different!
    • ​Yes they are! I cross-checked my values across a few different sources, though they do often vary by an order of magnitude.
    • For reference HarnManor has 6, GURPS  and Fief 15 and Allen 12 days of labor per acre of arable land.
    • The values for pastoral labor in Allen were harder to interpret. He uses 3.8 days for "commons, meadow and pasture", but also lists 41.8 for "meadow" and 14.9 for "improved pasture". As he doesn't specify what proportion of all fields was "improved pasture", calculating an average is hard. Using the lowest value gives a workable farm size that's close to the average size (15.1 vs. 14.1). That means that the more costly ones were less used.
    • (For what it's worth, Harn Manor also used 500 working days per household)
    • Allen had 262 for the farmer and 309 for the family, Apostolides et al. 381 for the whole family. Neither took into account child labor, which should be especially important at harvest when everyone was working in the fields. This is also where more data is required.
  • An average farm has only roughly 10 gp worth of animals and 10 gp worth of tools (one set of laborer's tools?)
    • That seems very low
    • Yes it does. I checked and that was based on the 10% annual expenses for capital. (10% of 80 gp/a times 33/12 for all capital = 22 gp). Then the question is what are the annual capital expenses?
      • 4 gp per year will get you a new set of tools every 2.25 years and a new ox every 10 years. As an ox has a two year working lifetime, that's too rare, but if we consider the capital expense as income loss from raising an ox instead of a cow to replace the old one, it's only 2.5 years per ox. Considering that not all farms had an ox (even though the 1 ox per two farms may be anecdotal) this will get us a realistic replacement rate.
    • What did you use as your conversion rate to calculate gp?
    • I just calculated the monetary value of production at ACKS! prices.
    • Dyer has an average of 1.7 horses/oxes, 2.9 cattle, 11.2 sheep, and 1.85 pigs per household
    • Broadberry has an average of 3.36 cattle, 42 sheep, 2.4 pigs per 30 acres
    • Allen has 1.36 horses/oxen, 1.67 cattle, 22.8 sheep and 0.95 swine per average farm (14+26 acres)
    • Apostolides et al. has 0.61 horses/oxen, 1.71 cattle, 13.66 sheep and 0.92 swine per average farm (14+30 acres)
    • Doubling the arable land size gets the values pretty close to Dyer and Broadberry
      • A rough conversion from livestock units to gp (10 gp per cow = 1 livestock unit) gives us 38.5 or 55.7 gp for Apostolides or Allen and 62.5, 78 and 75 gp for Dyer, Broadberry and ACKS!
    • If we calculate the capital from the total production, it's 220 gp.
      • Of that 40-60 gp is animals, 10 gp tools
      • If we add a hut for 25 gp, we're left with 135 gp for land and can say that the average English farmer could own 2.7 acres of his lands with ACKS! pricing
  •  Renting the land at the usual 1/33 per month is too expensive
    • ​Richard Cantillon indicates rents are typically between 1/3, 2/5, and 3/4 of the total production of the farm depending on the country. England had particularly low rents compared to other times and places.
    • True. From Allen the total for rent, tithes & taxes was 39% for 1300, 19% for 1500, and around 50% for 1700-1850

In any case, the nice thing about ACKS (at this point) is that all the assumptions are clear so you can change them to suit your particular time, place, productivity, etc. for your campaign.

Sure, but it was enlightening to see how the numbers match up when applied to one time and place. With more data we could have simple local modifiers for different eras, technological levels, cultures and climates.

To convert prices to ACKS, I typically do one of two things:

  • I start with an assumption that 1 silver penny = 1 silver piece (so a shilling is 1.2gp)
  • I inflation-adjust all prices to a fixed price of 4gp per quarter of wheat

(In some cases, ACKS prices and coinage are closer to ancient Roman values, but the English values, prices, and data is much easier to come by.)

Sounds easy enough. One more question: Is it assumed that most rural domain income comes from the peasantry and only urban domain income includes craftsmen, professionals and minor aristocracy? Or is there a percentage of the rural domain income that is from peasantry?

I'm asking because I'm trying to figure a way to keep the original domain income per family in a case like the 1300 era England. If land income is lower, could the money be extracted from the non-peasant classes?

The money in the hands of the urban families is going to be approximately equal to 1/2 to 2/3 the productivity of the peasant families. The reasons for that are that the lord or peasants will sell them surplus grain and use the money to buy goods from the urbanites.

I don't know if you've read Richard Cantillon (pre-classical economist) but he has a wonderful explanation of how the economy of a pre-industrial society functions, with a circular flow of the money in the society. I bulit an ACKS circular model on a similar basis. (For use in the next Axioms article, actually).

 

 

 

I have been thinking about the productivity multipliers for the craftsmen, but haven't really compiled good data yet. Since the cost of living is going to be the same for farm labor whether they work on their own or for a bigger landowner, withe the English model there's no extra income coming from the land. Too high population density (twice as many people on 30 acres farmland). The only ways to get more income would be other raw material extraction (mining, logging, hunting, fishing etc.) or on higher production by the urbanites and trade (eg. production of cloth for export from local wool). I'll have to find some data on those. Unfortunately the best dataset I have is from a later era. That and finding out th changes in agricultural productivity are bigger projects that I've been gathering materials for (actually for Traveller, but also for low tech fantasy world building).

I haven't read Cantillon, so I'm not familiar with that model.

Cantillon's book is available for free here. It's worth a read. 

https://mises.org/library/essay-economic-theory-0

 

Thanks! I'll check it out.