Notes on Stratification, Deprivation, and Quality of Life

Measuring Poverty: Two Approaches

Poverty is often equated with destitution, but other perspectives are possible.
Poverty gains historical significance when applied broadly, questioning if societies were generally 'impoverished' and how this impacted their development.
Two approaches to explore this issue:
- Asset and Income Distribution: Examine the overall distribution of assets and income in a society to determine how resource concentration at the top led to deprivation of the general population.
- Quality of Life/Human Development: Use quality of life or human development as an indicator of overall well-being. This considers non-economic factors influencing well-being.
Factors such as health, literacy, gender roles, and legal rights are easier to study than narrow definitions of poverty.
Enables cross-cultural comparisons within the ancient world and between ancient and modern societies and relates research to modern development studies.

Stratification and Inequality in the Roman Empire

Conventional imperial structure suggests a small elite ruling over a vast plebeian populace.
Imperial legislation favored the three orders (ordines) of senators, knights, and municipal decurions.
Distinction between honestiores (the three orders plus army veterans) and humiliores (pretty much everyone else).
Modern accounts of Roman social structure often use these legal categories as structuring principles.
Geza Alfoldy’s ‘‘St ¨ ande-Schichten’ (orders-strata) pyramid ¨ of Roman imperial society is an example.
Models derived from the ordo system or the honestiores/humiliores dichotomy are binary, separating the elite from the commoners.
Top-imposed binary ordering of society was not unique to the Roman empire.
- Similar to the distinction between elite and masses (khassa wa’umma) in early Islamic societies, like the Ottoman divide between askeri (political class) and re’ayya (common people).
- The Han empire in China had a more elaborate system with twenty ranks, where the lowest eight could be bestowed on commoners, extending stratification to the village level.
- These ranks provided exemptions from labor service and taxes, and sometimes land or office.
- Commoner ranks could be raised on special occasions, correlating rank with age.
- The top twelve ranks created a currency of honor and privilege for the official class.
- Only the top rank was hereditary.

Social Rankings and Economic Inequality

In imperial Rome, members of the three orders were generally wealthier than others.
Exclusion of rich freedmen shows that material wealth was a necessary but not sufficient condition for formal preferment.
The total number of honestiores was relatively small, approximately 1% of the empire's population.
- The three ordines consisted of at least 350,000 but probably not more than 500,000 individuals (including family members.)
- There cannot have been many more than 100,000 veterans.
The inclusion of veterans among the honestiores indicates that this group was not a homogeneous economic class.
The late Roman separation of potentiores from tenuiores was even less suitable as an economic marker.
Alfoldy argues that Roman imperial society lacked a genuine middle class, containing only upper and lower strata.
- He acknowledges that this divide did not strictly represent an economic dichotomy.
Andrik Abramenko’s attempt to identify seviri and augustales as a municipal middle class is problematic.
- These collegia formed a supplementary quasi-ordo, designed to accommodate wealthy individuals unable to join the local ordo decurionum.
- They represent a lateral extension of the third order, not a middling group.

The Imperial Ordines as Political Classes

Modern scholars identify the imperial ordines as political classes defined by their civic function.
In Runciman’s terminology, they are 'systacts', sharing a common endowment of power by virtue of their roles.
'Rulers' and 'ruled' may be an adequate rendering of ordines and plebs.
Garnsey stresses that honestiores and humiliores were not homogeneous groups, and that this dichotomy was largely confined to criminal law administration in Roman provinces.
Honestior was perhaps a functional category, lumping together the agents of the imperial center.

Dichotomous Models vs. Socio-Economic Reality

Karl Christ warns against taking dichotomous models as representative of socio-economic reality and stresses the existence of a middle stratum.
Vittinghoff considers it absurd to classify all humiliores as lower classes and denies the presence of an economic middle class in the Roman empire.
Current textbook wisdom tends to paint a rather different picture.
A leading survey asserts that while a sizable heterogeneous group of freeborn men existed, there was no middle class with independent economic resources or social standing.
Cruder versions of this binary view exist, where survival approximates the consumption patterns of the masses.
Peter Brunt finds no middle class in Rome, except for some rich freedmen.
Jerry Toner: 'There was no middle ground… The reality was nearer 99% poor, 0.4% military, 0.6% rich.'
German scholarship has adopted the same perspective, with middling individuals barely mentioned in surveys of the Roman imperial economy.
Such assessments leave little room for financially secure individuals independent of elite households but not wealthy enough for a leisured lifestyle.
The formal ordering of imperial society into a tiny elite and a vast humble mass seeps into evaluations of lived realities.
Willy Pleket pointed out that the image of the imperial plebs as a 'gray uniform mass' came into being as a foil for the 'honorable' people.
William Harris envisions an economic structuring of the Roman population into three classes:
- The well-to-do who relied on the work of others.
- Households that owned means of production but also engaged in work.
- Hired and slave laborers.

The 'Middle Class' in Roman Society

The lack of non-economic indicators obscures the size and significance of the middle class.
Epigraphic records of municipal handouts usually distinguish between decurions, civic associations, and an amorphous plebs or populus.
An inscription from Histria referring to gifts for carpenters and small businessmen is unusual.
Due to the agrarian character of Roman society, the allocation of resources among the rural population matters more.
Luuk de Ligt stresses the value of housing as an indicator of wealth distribution.
Tchalenko’s Syrian villages show a continuum from comfortable residences to humble shanties, rather than a stark polarity.
The distribution of garden land in the Fayum village of Karanis shows a smooth graduation from small to large owners.
The presence of goldsmiths in Roman Egyptian villages suggests local demand for luxury goods.
Overly dichotomized images of Roman imperial society are hard to reconcile with evidence from ancient Mediterranean civilizations, like the Greek poleis.
Archaic and classical Greece had a large socially and economically 'middling' population.
In classical Athens, Robin Osborne estimates that 7.5% of the population held about 30% of the land, while Lin Foxhall argues that 9% of households owned 35–40% of the land.
About 20 to 30% of the population may not have owned any land while 35 to 45% controlled about half.
Ian Morris observes that the Gini coefficient of 0.38–0.39 implied by these estimates is remarkably low by historical standards.
The apparent lack of very large estates in Attica is consistent with this reading.
The existence of a sizable hoplite 'middle class' is hardly in doubt.
The fact that the oligarchs of 411 bc could draw on 5,000 citizens with enough resources to equip themselves indicates the existence of a substantial ‘middle class’ (Thuc. 8.97).
At the apex of their prosperity at the beginning of the Peloponnesian War, the Athenians may have been able to muster no fewer than 22,000 adult male citizens of hoplite status (Hansen (1988) 24–5).

Athens: An Atypical Case?

The question remains of how typical Athens was in this regard.
Large-scale land allocations might produce even more egalitarian outcomes.
Oligarchic regimes may have exacerbated inequality.
Institutional arrangements were critical to specific outcomes.
Any weakening of the ‘middling’ ideology would facilitate land concentration in the hands of the few.
The Greek world benefited from widespread and significant improvements in living standards between 800 and 300 bc.
Conventional house size is one of the most powerful indicators: by the end of this period, homes were on average five times as large as at its beginning.
Other variables include longevity, stature, and nutritional status.
These improvements are traced to the gradual development of institutions that gave Greek citizens greater freedoms, specified property rights, and encouraged investment in human capital.
Factors include the poleis’ increasing freedom from predatory state rulers, the development of chattel slavery, citizen egalitarianism, the ‘invisible economy’ of banking, trade and commerce, and the relationship between war and economic growth.
This model undermines the notion that the Mediterranean had always been characterized by large-scale patronage, rural dependency, and extreme inequality in landownership.
Periods of economic growth temporarily outpacing demographic growth are repeatedly attested across pre-modern history.

Institutional Arrangements and Living Standards

Roman historians have yet to undertake similarly ambitious surveys of proxy variables of long-term economic growth and development.
The imperial success of the Roman state in the second half of the Republican period might have engendered comparable rates of per capita growth, at least in its core regions.
Differences in institutional arrangements may have militated against improvements in sub-elite living standards and favored increasing inequality.
Traces of widespread changes are common in the material record.
More synthetic and diachronic investigation will be necessary to establish the nature and scale of change over time.
If constraints on inequality are ultimately a function of civic and moral institutions, the Greek experience need not be particularly relevant to the Romans.
Even so, such institutional differences may be a matter of degree rather than substance.
For much of the Republican period, Roman citizen society was characterized by mass mobilization for war and participation in the political process.
The population was ordered in multiple formal categories that were based on timocratic principles.
Evidence for census rankings in the Republican period is a 'morass'.
Thresholds of HS100,000 for the first class, HS75,000 for the second, HS50,000 for the third, HS20,000 for the fourth, and HS375 for the fifth were assumed.
One could own as little as HS375 (less than the annual stipend of an infantryman) and still count as an assiduus.
The lower limit for the fourth class is more interesting, equivalent to about 5,700 to 6,700 modii, or 38 to 45 tons, of grain, translating to an annual income of between 1,900 and 2,700 kg of grain.
In a household of four, this yields 475 to 675 kg per person per year.
Peter Temin puts the average per capita GDP of the Roman empire at the equivalent of about 600 kg of grain.
The lower census limit for the fourth class appears to have approximated the income threshold for a reasonably secure commoner household.
Most members of the fourth class would have been reasonably well cushioned against chronic want.

Estimating the Relative Size of Roman Property Classes

The margins of error are very large but not infinite, and the formal structuring of the Roman citizenry must have followed some predictable pattern.
Despite the aristocratic focus of the Roman literary tradition, it is surprisingly hard to be more precise.
If we conservatively reckon with 10,000 knights in the late Republic, the ratio of voting centuries for equestrians and members of the first class suggests the existence of at least 40,000 citizens of the latter category.
If we stick, again very conservatively, to 40,000 first-class citizens and assume schematically that each of the remaining classes was one and a half times as large as the next-higher one, we get 60,000 members of the second class, 90,000 of the third, 135,000 of the fourth, 200,000 of the fifth, and 615,000 almost or completely propertyless citizens.
This model produces 225,000 third- and fourth-class households, out of a total of about 1.15 million households with 4 million citizens.
Under these circumstances, expropriation and pauperization would have been extremely widespread.
This model is seen as a limiting case that seems likely to understate the share of property holders in the general population.
It is hard to imagine how the number of middling property-owners could be smaller than the 20% share posited in this crude minimalist model.
What little empirical evidence we have is consistent with this model.
We need to allow for at least 20,000 decurions in early imperial Italy, many though by not necessarily all of them possessed of at least HS100,000.
In the alimentary land register of Ligures Baebiani in southern Italy, most estates fall in the range from HS25,000 to HS100,000, i.e. the census brackets for the second, third and fourth classes.
Spread out over more than 400 Italian cities and their territories, 225,000 third- and fourth-class households are sufficient to provide each of these communities with several hundred solidly ‘middle class’ families.
They would have provided the backbone sorely lacking from a hypothetical more extremely dichotomous society.
Collectively, they would also provide a mass market for moderately priced manufactured goods, cash crops such as wine, and even meat.
If these people existed in the late Republic, they cannot simply have vanished under the early monarchy, even if imperial sources show (even) less interest in sub-elite groups.
People were necessarily ignorant of the census divisions of a bygone age.

Asset Distribution in Egypt

Egypt is also the only part of the empire where surviving land registers and similar documents afford us a unique opportunity to trace patterns of inequality in asset distribution in some detail.
In a pioneering study, Alan Bowman computed a Gini coefficient of .815 for land owned by citizens of Antinoopolis and the residents of one of the four quarters of the city of Hermopolis in most of the Hermopolite nome in the mid-fourth century AD.
Bowman calculated a Gini coefficient of .737 for a list of private landowners in the Fayum village of Philadelphia in AD 216, a value which Roger Bagnall subsequently corrected to .532 (or .516 for complete datasets).
A list of crown tenants and cleruchs in the village of Kerkeosiris in the late Ptolemaic period (116/15 bc) produces a low Gini coefficient of .374.
In a later study, Bagnall used tax assessments to establish group-specific Gini coefficients for the village of Karanis in the Fayum in AD 308/9. The respective ratios are .638 for metropolitan landowners and .431 for villagers.
A later land register from Aphrodito in the Antaiopolite nome (c. AD 525/6) yields a Gini coefficient of .623.
All the Egyptian and other samples share a fundamental problem: they omit landless residents, many of whom were likely to be poorer than landowners.
Bagnall constructs a model for the Hermopolite nome that indicates an overall Gini of .56 for all landowners.
In this scenario, 59% of villagers’ holdings and 88% of those owned by urban residents were surplus to their personal subsistence requirements.
As many as 65% of all inhabitants of the nome could in theory have been landless.

Key Points from Bagnall's Model

Property was strongly concentrated among the top 10% of all landowners.
The model envisions 952 local urban landowners, which translates to an elite segment of about 100 individuals.
There were few if any large landowners outside a council-sized group.
The model also generates 7,400 rural landowners, 59% of whom owned enough land to enjoy a net surplus.
Bagnall refers to them as ‘a broad band of middle-range men capable of bearing public obligations’.
At least 20 to 25% of the entire nome population belonged to the 'middle class'.
The concentration of a large percentage of all assets in the top 3 to 5% of the population is compatible with the existence of a substantial middling group of owner-producers.
The Hermopolite model echoes the earlier guesstimate that at least 20% of Italian households belonged to the third and fourth census classes.
There is sufficient evidence in support of the notion of an economic continuum from a narrow elite to a steadily broadening middling group.
It is perfectly possible to reconcile the dominance of a disproportionately affluent elite with the presence of a substantial ‘middle’.
The relative size of the ‘poor’ population, is of course much more difficult to pin down.
Whether the majority of the nome population really lived at a very low level of subsistence is what we would need to establish rather than merely assume.
The key question is among how many consumers the available surplus was distributed.

The Quality of Life

How are ancient historians supposed to measure living standards anyway?
Per capita GDP would be a measure of dubious value, and not just because it is unknown.
Observing that urbanization and GDP tend to be correlated in modern developing countries, Temin argues that likely levels of urbanization in the Roman empire imply an average per capita GDP equivalent to about $2,000, comparable to that of India.
Even if this were in fact a meaningful parallel, what would it tell us about Roman living conditions?
In many ways, India is not at all like the Roman empire.
Dissatisfaction with this conventional measure inspired the launch of the annual Human Development Report, whose goal is to devise a series of indices that take account of a broad variety of factors that impact upon living standards, including income, demographic indicators, health, education, literacy and school enrolment, inequality in income and consumption, priorities in public spending, unemployment, energy consumption, refugee displacement, crime, gender empowerment and inequality in education, economic activity and political participation, and human and labor rights.
There is no good reason to assume that the ancient Mediterranean was much more homogeneous than, say, sub-Saharan Africa or Latin America today.
Bruce Frier, in a pioneering attempt to assess the quality of life in the Roman empire, looks at various factors only to pass an unfavorable verdict – life was short, literacy rare, and so forth, which means that the Roman state did little to improve the lot of its subjects – but fails to specify his criteria of judgement and standards of comparison.
Ian Morris focuses more narrowly on consumption (of material goods) by considering a whole bundle of variables that permits us to get a rough idea of the scale of change in overall consumption between 800 and 300 bc.

Measuring Deprivation

The study of human well-being and deprivation in the ancient world needs to be separated from the study of economic growth.
Attempts to measure deprivation are complicated by the fact that it cannot be measured independently but is contingent on social standards that define poverty thresholds.
Sen expanded his entitlement approach by turning his attention to the influence on well-being of political freedoms, economic facilities, social opportunities, transparency guarantees, and protective security.
It would be a worthwhile exercise to re-configure and re-view our evidence for different ancient societies to explore, in a systematic fashion, the characteristics of each of these factors and their change over time.
Our ultimate goal has to be a comparative evaluation of different ‘poverties’.