The Economics of Roman Slavery - Kyle Harper

Amazon link

After reading Kyle Harper's excellent "The Fate of Rome" I was motivated to read this earlier book (an edited version of his doctoral thesis) on the nature of slavery in the late empire. I'm some way into it and the book is filled with astonishing detail.

Harper brings an economist's eye to modelling late Antiquity. Some of his neoclassical micro- and macroeconomic concepts are jarringly anachronistic - but despite his criticisms of dogmatic, mechanistic and dated Marxist analyses of Roman slavery his empirically-based research is very assimilable to a more sophisticated Marxist approach. Yes, Roman society integrated multiple modes of production into its interlaced agrarian-mercantile economy.

In Harper's working model, the late Roman empire comprises around 50 million people, of whom 5 million (10%) are slaves. The bulk of the slaves are owned by the top 600 or so senatorial families and the next 1.3% of super-rich aristocratic families beneath them. But the next 10%, professional/artisanal households in the cities and landowner holdings in the countryside, would have owned a few slaves (1-5) each. The remaining 80% of free citizens are the (mostly agrarian) poor working their little subsistence (often leased) plots.

There was little free labour in the western empire: population densities were low and there were other occupational calls for poor freemen, particularly the military. In Egypt, where poor labourers were in abundance, wage labour was much more central.

Slavery endured over a number of centuries. Routine supply-demand calculations show naive assumptions that the supply of slaves was provided solely, or mostly, by war captives are simply untenable. Slaves were expensive and valuable pieces of capital equipment - with the additional utility that female slaves could produce more of them. Slaves were bred just as cattle were. Slave families were a thing.

From pages 150-151:

"Scheidel argues that slave labor emerged in classical Athens because it was cheap while wages were high and that similar conditions perhaps existed in the late Roman republic.

In the empire, by contrast, slaves were dear but their labor was profitable, so that slavery persisted in a high-equilibrium state. Indeed his model is a corrective against one of the most persistent illusions about slave labor: that slaves are necessarily cheap labor.

What matters is the ability to employ slave labor at a profit  - the marginal value of labor against the costs of a slave, which are often high. The emphasis on supply is necessary and even obvious; no critique of the conquest thesis should deny that the supply side was always an economic constraint, as important as the demand side.

But the dynamics of demand are, analytically, more interesting and more challenging. Scheidel's model begins by proposing that slave labor and free labor are in more or less direct competition. Landowners used slave labor not just because free labor was institutionally unavailable, as per Finley, but because it was expensive. It was expensive because it was in short supply. It was in short supply due to high levels of political/military commitment among the free population or because of low population levels.

This model explains why slave labor is so prominent in late republican Italy (with its military mobilization) and so exiguous in Roman Egypt (with its dense population). Scheidel's model adds an important variable complicating the proposition that free and slave labor were in perfect competition in some sectors of agriculture.

He adopts Hanes' arguments about the role of transaction costs in the employment of slave labor. The risk of being unable to find labor at sensitive times in the agricultural cycle, the time and effort spent contracting labor, and the turnover costs of retraining workers were all salient considerations. Slave labor was effectively an involuntary lifetime contract.

Transaction costs, then, should be accounted for along with the raw market price of wage labor. Transaction costs amplified the effects of the demographic context, for "thick" labor markets like those in Egypt not only drove down wages, they reduced the risks and costs of the open labor market for land-owners.

This model also explains why slave labor was adaptive to certain types of Mediterranean agriculture, such as grape and olive cultivation, that required both effort and care in the nature of the labor performed. It thus represents an important advance in the application of institutional economics to ancient slavery by taking seriously the nature of the work regime and the type of labor employed. "

I wouldn't recommend the book as a beach read, it's dense and academic. But it's marvellously informative and raises this question: if we ever develop the robotic analogues of slaves, would our society evolve into anything like that of the late empire?