Scheidel Walter 2005 ‘The comparative economics of slavery in the Greco-Roman world’, Princeton/Stanford Working Papers in Classics, Nov. ’05, 21p.
Large scale slavery occurs rarely in history, Ancient Greece and Italy are exceptions. Why? The slave population in Athens and Rome can be very roughly estimated at about a third of the total, a proportion consistent with the one of ante bellum Southern USA (2).
Why slaves and not waged labour? Theoretically, slaves can be used in capital- and/or care-intensive activities (some fragile crops, commerce, artisanal duties) or pain- and/or effort-intensive ones (mining). The former require rewards and incentives and the latter close and often harsh supervision (3). Slavery seems particularly indicated for activities requiring little skills and easy to supervise.
But ancient slavery does not fit easily in this model. For viticulture, chain-gangs were commonly used in Italy despite its care-intensive character. Similarly, the use of slave in services in Rome is inconsistent with a restriction of slavery to unpleasant tasks inspired by 19th century USA (5).
One explanation for the comparative advantage of slavery during the Antic period is that, in “thin” labour market conditions, in an agrarian economy where 5 days can make the difference between a successful harvest and a failed one, the “turnover” cost (transaction cost including searching a replacement for a job-specific worker and the necessary time for a worker to adapt to his new workplace) can be enormous.
In these conditions, slavery stabilizes the workforce and prevents turnover costs. This theory is supported when Italy and Greece are compared to Egypt which had a “thick” labour market (high density, low wages) and few slaves.
The use of slave craft production in otherwise thick urban labour markets during the Antiquity can be explained by the “openness” of ancient slavery (high chances for a slave to be freed compared to the “closed” race-based US system). Besides the prospect of being freed was certainly rewarding enough to support the use of slaves in care-intensive industries (6).
The Roman system and to a lesser extend the Greek one let the slave owner the ability to pick the system he thought the most efficient: reward for care-intensive activities or supervision for effort-intensive ones (8 ).
Conditions for the development of slavery
For slavery to be an effective solution, the following conditions are necessary: high real wages for free labour, demand for goods that can be produced by slaves, access to slaves and significant capital accumulation (to buy the slaves and consume their production). These conditions were all present in the USA, But labour scarcity is not present in the central areas of both Greece and Italy. (p.9) In that case, labour scarcity may have been caused by the political, judiciary and military requirements of the polis life. Thus victory and democracy made slavery more likely (10).
Costs and choices
Athenian slave prices were significantly lower than in Rome, and the attractiveness of slavery must have decreased under Augustus. Large scale slavery may have disappeared from some sectors of the Italian economy even though a pro-slavery cultural path dependency may have by then been established. The Antic World lacked the shocks that ended slavery in the 19th century. Once created, the large scale slavery just requires a steady supply and stable price to be maintain. “In consequence, the core of the Roman Empire may have entered a prolonged ‘high-equilibrium’ state of slave-holding” (15).
Consequently, slavery may not have been the most efficient production system over most of the period, but in the real world, where information is costly, it may have been based mostly on tradition, cultural prejudices and other socio-anthropological trends (16).
Several aspects remain unaccounted for in this article. The first one – lets follow the comparative perspective of the author – is that none of the characteristics described of Athens and Rome when slavery became prevalent were unknown, say, in 16th century Spain. So why Rome and not Spain? Overall, all the comparative advantages of slavery can only kick in under some specific conditions that have not been analysed in this article.
Another issue is the qualification of the classical Greek and Roman labour markets as ‘thin’. They were by modern standard but a great many other societies had thinner labour markets and did not develop slavery. Greece and Rome both exported a lot of men in military and colonial adventures which is hardly consistent with thin a labour market. Potentially, the thinness in these cases came from a lack of fluidity in the free labour market (cast-like system, prejudice against some activities, preference for family-based business model) rather than from a demographic deficit as the competition between free and servile workforce in Sicily made clear.
A comprehensive modelling of classical slavery remains to be done.
Fenoaltea, Stefano 1984 ‘Slavery and supervision in comparative perspective: a model.’ Journal of EconomicHistory 44:635-68
Hanes, Christopher 1996 ‘Turnover cost and the distribution of slave labor in Anglo-America.’ Journal of Economic History 56:307-29