by Ober, Josiah
At the other end of the time period, Morris’ estimate of the population of the Greek world in ca. 300 BCE is in line with other demographic estimates since the nineteenth century. Moreover, it is substantially lower than the more recent, detailed, and highly plausible estimates of the Danish historian, Mogens Hansen, who uses different estimation methods, addresses a substantially larger geographic area, and includes communities that were incompletely Greek in culture. My own model of Greek demographic growth, based on Hansen’s estimates, is illustrated in figure 2.1. Morris’ claim of a tenfold increase should be regarded as a minimum.16
If we accept Hansen’s mid-range figure of more than 8 million for the total population of the world of the poleis in the fourth century, meaning that (as well as accepting the modeling technique) we define “Greek” as a person living in a community that was substantially Greek in culture (rather than “Greek by ancestry”), it appears that the Greek population increased by a factor of 25 in the 700-year period between the Early Iron Age nadir and the peak of the classical efflorescence.17 The geographic expansion of the Greek world accounts for more than half the rise in total population; see figure 2.1. The figures for “core Greece,” discussed below, suggest a nadir-to-peak local population density increase in the range of eight- to tenfold, but this discounts all colonization and emigration.
The second key factor in estimating aggregate economic growth is per capita consumption. Morris sought to estimate changes in per capita consumption over the same 500-year period, 800–300 BCE. While there is no way to measure consumption directly, the proxies employed by Morris are telling. Morris assembled a substantial data set (n = 405) of Greek house plans. The median Greek house in the ninth century was small and squalid. Over the next 500 years, the median house became both much bigger and much better built. Looking at square footage alone, when account is taken of probable second stories, the change of the median house is more than 350%—from ca. 80 m2 to ca. 360 m2. Given the striking improvement in building standards, the total increase in the economic value of a house would actually have been substantially greater. Morris notes the difficulty of measuring the change in other consumption goods, but based on archaeological evidence of sites destroyed suddenly, he posits that, over the period 800–300, “a five- to ten-fold increase … seems reasonable.”18
Moving from these numbers to total per capita consumption is a complex problem; a big part of premodern consumption was in the form of food and (where applicable) taxes and rents. Morris argues, on very reasonable grounds, that per capita consumption in ninth century Hellas must have been close to the subsistence minimum. By 300 BCE, however, he suggests that consumption had increased by at least 50% and perhaps as much as 95%. Thus, by 300 BCE, a typical Greek household was consuming half again to twice what an ordinary household had been consuming 500 years before. This range yields a per annum growth rate in per capita consumption of 0.07–0.14%. By comparison, the growth in the Roman per capita growth rate has been estimated at 0.1%.19 I return to the question of Greek per capita growth below, showing that Morris’ upper-range estimate is more likely than any lower estimate and that the actual rate of Greek per capita growth during 800–300 BCE was probably about 0.15%—one and a half times the estimated Roman growth rate.
Combining his estimate of demographic growth with his estimated growth in per capita consumption, Morris posits that total aggregate consumption growth (number of people × rate of consumption) in Hellas increased roughly 15-fold (assuming his lower per capita rate) to 20-fold (assuming the higher per capita rate) in the period 800–300 BCE, for an annual aggregate economic growth rate of 0.6–0.9%. As Morris points out, Holland is the gold standard for a high-performing early modern economy. The annual aggregate growth rate for Holland in 1580–1820 was about 0.5%. And so, as Morris notes, even if we were to cut his estimate of growth in half (and, per above, we have no reason to think we ought to do that), the Greek economy compares favorably with an exceptionally high-performing premodern economy.20
Morris’ conclusions about relatively high per capita and aggregate Greek economic growth are consistent with other indirect proxies that point to substantial growth in the late archaic and classical periods. Based on data made available in the on-line version of the Oxford-based Lexicon of Greek Personal Names, I calculate that the number of known (from literary or archaeological sources) names of people in Attica (the territory of Athens) grew from ca. 1,200 in the sixth century to ca. 17,000 in the fourth century, an approximately 14-fold increase in less than 300 years. The increased visibility of individuals is obviously the product of multiple variables, notably growing rates of literary and epigraphic production. This high rate of growth in name visibility is consistent with a world in which many more people were consuming substantially more. Counterfactually, a world with a declining population in which most people lived at a level of bare subsistence would clearly be less conducive to rapid growth in the literary and epigraphic visibility of people’s names.21
Based on data taken from the Inventory of Greek Coin Hoards, David Teegarden and I estimate (in an unpublished study) that the volume of coined money circulating in the Greek world increased substantially as well. between the sixth and fourth centuries BCE, the median size of a Greek coin hoard (an indirect proxy for per capita rather than aggregate growth) roughly doubled, from 23 coins to 48 coins per hoard. Meanwhile, the average hoard size quadrupled, from 52 coins to 213 coins, reflecting the increasing incidence of some exceptionally large hoards.22
When looking at the total number of hoards, and at the total number of coins in all known hoards (which ought to be indicative of aggregate growth), the sixth century numbers are misleading since coinage was introduced in the Greek world in the course of that century. Yet even when we restrict our survey to the classical period, the numbers are suggestive. The number of hoards more than doubled from the fifth century to the fourth, from 238 to 564 hoards, while the number of total coins in all hoards grew threefold, from about 34,000 coins to about 109,000. These numbers cannot readily be translated into a given annual growth rate. Short-term growth in hoarding may, in fact, indicate economic crisis.23 But over the long term, the substantial growth in both hoard size and numbers of hoards is likely to reflect a world in which there was more money in circulation and in which more people could afford to save some part of their income in the form of cash. This ought, in turn, to mean a world in which more people were living substantially above the level of bare survival.
TABLE 4.1 Summary of Proxy Indicators of Economic Growth
NOTE: Dates represent centuries BCE.
Table 4.1 sums up the evidence for change over time in the proxy indicators discussed in this section: They all move in the same positive direction; they all point to substantial growth over time in the classical Greek economy.
DENSE, HEALTHY, URBANIZED POPULATION
In addition to the relative measures of growth over time, we can evaluate population and welfare in a given era in absolute terms and thereby compare societies over time and space. Denser populations and high levels of urbanization tend to be correlated with higher economic performance and are, therefore, commonly employed by economic historians as proxies for economic growth.24 Among the important results achieved by the Copenhagen Polis Center, directed by Mogens Hansen, has been to give us a better sense of the total population of the extended Greek world in the late classical period and the distribution of that population. As we have seen (ch. 2), by the later fourth century, there were an estimated 1,100 poleis in the Greek world. The territorial size of about two-thirds of all documented poleis (672/1,035) is now known or can be plausibly approximated.
Beginning with the extensive empirical evidence of his inventory of poleis, Hansen uses what he calls a “shotgun method” to estimate late fourth century Greek population and distribution. The method employs the evidence of the physical size and estimated population densities of relatively well-documented poleis, along with the size and estima
ted densities of intramural areas and the known distribution of poleis across a range of sizes, to arrive at overall estimates of population. On the basis of his shotgun method, and in comparison with 22 poleis for which there is literary and documentary evidence for individual polis populations, Hansen offers new estimates for the minimum total population of the Greek world, the distribution of the Greek population among large, middling, and small poleis, and the relative scale of urban and extraurban populations.25
It is based on this method that Hansen estimates that the extended Greek world in the late fourth century BCE had a population of at least 7.5 million, and more probably 8.5–9.5 million people. The extended Greek world considered in Hansen’s estimate includes regions of Greek settlement excluded from Ian Morris’ estimate of ca. 4 million for the Aegean, Sicily, and southern Italy, but Hansen also argues for a somewhat denser population overall. As we have seen, assuming a late fourth century population of 8.25 million means that the number of Greeks, taken in toto, had—through a combination of natural growth, immigration, enslavement, and the Hellenization of formerly “barbarian” towns—multiplied by a factor of about 25 in 700 years, from the Early Iron Age nadir in ca 1000 BCE to the classical peak in ca. 300 BCE. As noted in chapter 2, using these figures we can estimate that by the late fourth century the Greek world had an overall population density of ca. 44 people per km2—very similar to that of Holland in the mid-sixteenth century and to England and Wales in the late seventeenth century.26
Hansen argues that what I am calling “core Greece,” the parts of Hellas defined by the extent of the Greek state in the later nineteenth century, had a population of at least 3–3.5 million people.27 By way of comparison, in 1889 a census by the state of Greece counted 2.2 million people—this should probably be corrected upward to 2.3 million, per discussion at the beginning of this chapter. As Hansen points out, and as is confirmed by historical demographer Vasilios Valaoras’ analysis of the causes of Greek unemployment and large-scale emigration, there is reason to believe that the population of Greece in the late nineteenth century had already exceeded Greece’s agricultural carrying capacity. The land of Greece simply could not produce enough food to feed any more people. If these figures are in the right range and if the assumptions about carrying capacity are correct, it has considerable bearing on the performance of the ancient Greek economy.28
Unless we are willing to assume that fourth century BCE Greece was much more agriculturally productive than nineteenth century CE Greece (which seems on the face of it unlikely—although it would be interesting if true), if we adopt Hansen’s figures, we must suppose that a substantial part of the fourth century Greek mainland population was fed from food imported from abroad. Something like 0.7–1.2 million Greeks, i.e., roughly a quarter to a third of core Greece’s population in the fourth century BCE, thus may have lived on grain imported (e.g.) from the western Mediterranean, from the Bosphorus/Crimea, or from North Africa.
Once again, even if we were to cut the number of Greeks (derived from Hansen’s estimates) who must be presumed to have lived on imported food in half, we would still be left with some 350,000–600,000 people in excess of the presumed carrying capacity of core Greece.29 This means, in turn, that Athens cannot have been the only major grain-importing Greek polis. And so, the core Greek world can no longer be regarded as entirely defined by subsistence agriculture or local exchange. The imported food had to be paid for somehow—by commodity exports (oil, wine, silver), manufactured goods, services, or the extraction of rents (i.e., by the use of power to obtain resources at prices lower than those that would pertain in a competitive market).30 The general point is that by the fourth century BCE, the mainland Greek world evidently pushed back the standard premodern limiting factor: the “low Malthusian ceiling” of resource constraints.
Total population is only one part of the equation. If we are to understand the conditions of Greek economic growth, it is important to determine how the population was distributed. My calculations, based on Hansen’s analysis of the numbers of large, middling, and small poleis that ever existed in the Greek world shows that most poleis (ca. 900 of ca. 1,100) were small (sizes 1–3), with populations of ca. 1,000–7,000 people. Yet only about a third (36%) of the total Greek population lived in these small poleis. Another quarter of the population (26%) lived in fairly large poleis, that is, in communities with a median population of about 17,000 people (size 4). About 4 in 10 (38%) of the polis-dwelling Greeks lived in very large poleis (sizes 5–7), with a median population of about 35,000 people or more. The results are tabulated in table 4.2.31
Along with population density, the rate and level of urbanization are commonly used by economic historians of premodernity as a proxy for economic growth. Among Mogens Hansen’s most striking claims, based once again on the shotgun estimation method, is that about half of the population of late fourth century Greece lived in intramural “urban” centers.32 Combined with the distribution into small, middling, and large poleis, this suggests that about a third (32%) of all Greeks lived in towns of 5,000 people or more: the standard for “urban” that is used by many demographers studying premodernity. A 32% urbanization rate means that, across the Greek world, about 2.5–3 million Greeks lived in what modern demographers define as cities. Assuming that these estimates are in the right ballpark, the Greek world of the fourth century BCE had a much higher urbanization rate, overall, than the Roman imperial world of the first and second centuries CE, in which some 10–12% of the population (7–8.5 million people) lived in similarly large towns. Rome’s urbanization rate was roughly similar to that of England and Wales in the seventeenth century, or France in the eighteenth century.33 Classical Hellas was less urbanized than mid-seventeenth century Holland (45%); but England and Wales reached classical Greek urbanization levels (and a similar number of total of urban dwellers) only in the first years of the nineteenth century (table 4.3). Modern Greece did not equal the fourth century BCE urbanization rate until the 1930s (figure 4.1). The picture does not change significantly when we use a more demanding standard of urbanization.34
TABLE 4.2 Estimated Distribution of Greek Population by Polis Size
NOTE: Based on ranking of polis size distribution modeled in table 2.1.
The high urban population of classical Hellas fits comfortably with the conclusion that a good many Greeks living in the core areas of Hellas (as defined by the frontiers of the late nineteenth century Greek state) consumed imported rather than locally grown food. It would be wrong to imagine that the set “urban residents” completely overlapped with the set “imported food consumers”; a substantial number of urban Greeks may have lived in “agro-towns” from which residents commuted to their fields. Some very extensive intramural areas may have enclosed gardens or even fields. Nevertheless, the estimate that about a third of the Greek world lived in urban areas is compatible with the finding that roughly a quarter to a third of the core Greek population was fed from imported food. Both results push against the standard modern premise, which assumes that the ancient Greek economy was overwhelmingly defined by subsistence agriculture; together the two demographic results point to a relatively sophisticated and diversified economy, one in which many people lived well above bare subsistence. These population figures are, in turn, in line with recent work, notably by the Chicago-based economic historian, Alain Bresson, to the effect that trade, in commodities as well as luxury goods, was much more important in the Greek economy than was long thought to be the case.35
TABLE 4.3 Comparative Urbanization Levels and Populations
Urban %
Urban Total (millions)
Hellas 350–300 BCE
32
2.5–3
Rome 100–200 CE
10–12
7–8.5
Holland 1651
45
0.44
England and Wales 1688
13
0.74
France 1788
12<
br />
2.8
England and Wales 1801–1803
30
2.78
NOTES: Urban = Town of more than 5,000 people. Figures for Rome: Wilson 2011. Figures for Holland, England and Wales, and France: Milanovic, Lindert, and Williamson 2011: Table 1.
Higher levels of urbanization correlate, historically, with higher incomes and economic intensification,36 but not necessarily with improved health and welfare: Rapid growth of urban populations has historically been associated with the spread of disease, and, e.g., in nineteenth century England and Holland, with squalid living conditions in crowded tenements. There is no evidence that these dismal conditions pertained in fourth century Greek towns. While the data on change over time in the health of Greek populations are difficult to interpret, and in some ways contradictory, it is clear enough from studies of human bones found in Greek archaeological excavations that the average life span of Greek men and women reaching adulthood increased substantially from the end of the Dark Age to the fourth century BCE. Based on the more recent (1990s) analyses, the ages at death of individuals surviving childhood seem to have increased by about 10 years for both men and women over this period: from about 26 to 36 for women, and from under 30 to about 40 for men.37
Life expectancy at birth (which in all premodern populations is much lower than average age at death for those surviving into adulthood, due to high levels of infant and child mortality and to the effects of disease) in fourth century Greece would still have been very low by modern standards, perhaps not exceeding the mid- to upper twenties.38 But it was certainly substantially better than it had been 500 years previously, in Early Iron Age Greece, when there was probably no town of as many as 5,000 people. Moreover, despite intensive archaeological exploration, there is not as yet any evidence in excavations of classical Greek towns for extensive tracts of small and squalid urban dwellings. As we saw above, the median Greek house, urban as well as rural, tended to become much larger and better built in the five hundred years after 800 BCE. Given the intensity of the archaeological exploration of Greece, it seems unlikely that slums have simply escaped notice.39