Why did the Roman Economy Decline?
I want to reflect on one leading account for the economic decline of Europe following the collapse of the Western Roman Empire. I recently encountered an explanation of this decline that strikes me as deeply problematic.
This argument is worth paying attention to as it is advanced by Peter Brown of Princeton University, who is among my favorite historians of late antiquity, and the author of influential, insightful, and often beautifully written reflections on religion and society in the late antique world. As a writer, he has the ability to make the ancient world come alive in original and unexpected ways. I particularly admire his biography of Saint Augustine and recent work on wealth in early Christianity. Nevertheless, in recent work he has been advancing a particular explanation of the decline of the Roman empire which strikes me as incompatible with both basic economics and what we know about other comparable preindustrial societies.
I’ll focus on the summary of the argument that he presents in The Rise of Western Christendom (I’ve been reading the 2013 Tenth Anniversary Revised Edition). His presentation draws heavily on Christopher Wickham’s Framing the Middle Ages . But I focus on Brown’s version here.
It should be clear that I am writing this as an economist and not as an ancient historian. But the problem with Brown’s account does not lie with his treatment of evidence or mastery of the source material but in his use and misuse of economic concepts.
The Roman State as the Engine of Growth?
In the Rise of Western Christendom, Brown summarizes the new wisdom on the transition from late antiquity to the early middle ages. He accepts that this transition brought about an economic decline — a decline evident in the radical simplification in economic life that took place. Long distance trade contracted. Cities shrank and emptied out. The division of labor became less complex. Many professions common in the Roman world disappeared.
All of this is relatively uncontroversial. At issue is what caused this decline? Traditional accounts emphasized the destruction brought about by barbarian invasions and civil wars as the frontiers of the Western Empire collapsed. These accounts emphasized a collapse in trade and increased economic insecurity. Brown, however, argues that the bulk of modern research rejects this old fashioned view. Instead, according to Brown:
The fault lay with the weakening of the late Roman state. The state had been built up to an unparalleled level in order to survive the crisis of the third century. The “downsizing” of this state, in the course of of the fifth century, destroyed the “command economy” on which the provinces had become dependent (Brown 2013, 12).
The barbarian invasions, of course, play a role in this story because they put pressure on the Roman state. But their role is peripheral. Rather, Brown contends that the Roman state was the engine of economic growth of late antiquity. Turning on its head the old view associated with Michael Rostovtzeff that attributed the decline of the Roman economy to high taxes imposed by the Emperor Diocletian and his successors, Brown argues that these high taxes were in fact the source of economic dynamism:
High taxation did not ruin the populations of the empire. Rather, high tax demands primed the pump for a century of hectic economic growth. Fiscal pressure forced open the closed economies of the countryside. The peasantry had to increase production so as to earn the money with which to pay taxes” (Brown, 2013, xxv)
He unabashedly presents the state as key to the late Roman economy:
With its insistent gathering of wealth and goods through taxes and their distribution for the maintenance of large armies, of privileged cities, of imperial palaces, and of an entire ruling class implicated in the imperial system, the late Roman state was the crude but vigorous pump which had entered the circulation of goods in an otherwise primitive economy. When this pump was removed (as in Britain) or had lost the will to tax (as in Merovingian Gaul and in the other “barbarian” kingdoms) the Roman-style economy collapsed (Brown 2013, 13).
This, I should add, is not presented by Brown as a tentative hypothesis or conjecture, but introduced as the current historical consensus. Brown is damning of historians who deviate from it, and particularly contemptuous of those inane enough to blame the decline on invading hordes of Germanic barbarians.
Let us grant that Brown is correct to present this argument as the current consensus among historians of late antiquity. The problem with it is that it is at odds with what standard economics and with what economic historians know about other preindustrial societies. To see why it is so flawed, I’ve done my best to reconstruct the argument.
- The first premise of Brown’s argument is that the Roman state was a sufficiently large player in Roman economy, in terms of the taxes it collected, and the money it spent on wages and armaments, that a reduction in state expenditure would have had a major impact on the rest of the Roman economy. And that any reduction in state spending would not have been compensated for by an increase in private spending.
- The second critical premise in Brown’s argument is that, in the absence of the demands of the tax collector, peasants would not have participated in the market economy. When “the great engine of enrichment stalled and, eventually stopped,” Brown writes: “No longer disciplined by the tax collector, the peasantry slacked off. They returned to subsistence farming”. It is crucial for Brown’s thesis that, without the pressures of the state, peasants would have produced only for subsistence.
- The third premise is that urban economy of the Roman empire served solely or predominantly to satisfy demands of Roman elites whose incomes were crucially dependent on the state. So when these state incomes declined so did Roman cities.
From these three premises, it follows that when the ability of the Roman state to collect taxes and spend tax revenues became severely damaged in the fifth century, the Roman economy went into fairly rapid decline. No longer forced to pay taxes in cash, peasants ceased producing goods for market. No longer emporia for the disbursement of state largesse, the cities of the western Empire went into decline. As the Roman fiscal state declined:
incomes dwindled, the rich no long reached out, as they had done in the glory days of the fourth century, to buy fine pottery, statuary, high-quality wines and exotic foods. They made do with the products of their region.
This, then, is Brown’s explanation for the decline of the Roman economy. It turns out that when examined one by one each one of these premises is either on shaky grounds factually, economically, or requires us to make implausible assumptions.
I think the first premise rests on a misunderstanding of textbook Keynesianism. Simply put, conventional Keynesian theory suggests that in a recession when resources are unemployed, an increase in government spending can in the short-run increase aggregate demand (either directly or via inflationary expectations). The details of this simple proposition have been endlessly critiqued and debated. But we will skip over this. What is important to note is that for standard textbook Keynesianism, this is a short-run effect. In the textbook models aggregate demand should eventually recover (via the real-balance effect). Government spending has the ability to speed up the recovery.
None of this suggests that in the long-run government spending is required to “prime the pump”. Indeed this language suggests a misunderstanding. For conventional Keynesians, the multiplier on government spending boosts short-run aggregate demand, but aggregate demand is not the binding constraint on long-run growth, supply is; growth depends on the productive capacity of the economy.
If anything, the impact of the Roman tax state on the productive capacity of the economy was more likely to be negative rather than positive. Resources were diverted from the private hands of peasants, merchants and small landowners and diverted into the hands of soldiers and officeholders.
For Brown’s thesis to hold, therefore, the Roman economy must have been in danger of continuous secular stagnation. Brown’s second premise alludes to one such source of stagnation. If peasants refused to participate in the monied economy this could indeed be a source of involution. That is, if peasant incomes went up and they spent none of this on urban-based manufactured goods but consumed the entirety of their higher incomes in the form of greater leisure. That is, Brown’s argument requires that at the margin, peasants preferred additional leisure to the wide array of affordable manufactured consumers goods that were on offer in markets and shops across the Roman empire. This is not impossible. But it is at odds with what we know about peasant behavior in other commercial societies such as early modern Europe. If we relax this highly implausible assumption, then the argument that the urban economy required the fiscal-military state, falls apart.
A similar knife-edge assumption is required for his third premise. Research on the earlier Roman empire suggests that the large cities of the empire were not merely “consumer cities” parasitical on the countryside but centers of urban production and manufacturing (see). Brown’s argument requires us to believe that if, for instance, the Roman state stopped spending on armor and weapons in a city, then the blacksmith and armor manufacturer would go out of business. This is a classic case of focusing on the seen and missing the unseen. It neglects the fact that lower taxes would give individuals more disposable income and they would likely spend some of this income to purchase amphora, pottery, textiles or other urban goods that we know the Roman economy was capable of producing. The blacksmith might switch to producing pots and pans rather than swords but he would not then go out of business.
If Brown’s account is implausible, what does account for the decline? The best account I am aware of is Bryan Ward-Perkin’s The Fall of the Roman Empire and the End of Civilization which Brown dismisses as “tendentious and ill-supported polemics”.
The collapse of the Roman state was catastrophic, not because the Roman state was an engine of economic growth, as Brown contends, but because it provided, albeit imperfectly, the public good of defense. In the absence of this, transactions costs greatly increased, long-distance trade declined, markets contracted, and urbanization declined.
Addendum: The Size of the Late Roman state.
Here I explore a related aspect of Brown’s account that appears problematic: his claims about the size of the late Roman state.
Brown claims that the post-Diocletian Roman state was a “command economy” capable of mobilizing tremendous resources and driving the Roman economy. I am skeptical of such a description being an accurate description of a premodern state. Prior to the railway and telegraph, there were severe limits to ability of states to direct economic activity. To get a feel for things I tried a back of the envelope calculation of the size of late Roman fiscal military state.
The main component of the Roman state was the army. The army grew considerably after Diocletian’s reforms. The exact size of the new army is subject to considerable controversy. John Lydos estimated the Roman army to comprise 389,704 men and a navy of 45,562. The largest estimates are based on Agathias and date from the mid-sixth century. These suggest that the total size of army and navy was around 645,000 (580,000 in the army and around 65,000 in the navy). Historians tend to think these number are too high but we will accept them for the purpose of the argument (Agathias is critiquing the government of his day by showing that the army had declined greatly from the days of Diocletian).
Unfortunately estimates of the Roman population are extremely rough and we don’t have any good numbers of the 3rd century. The population of the Roman empire c. 160 is estimated to have been between 60–70 million. The population in 300 was likely lower than this, though it is unlikely that it was substantially smaller, as recent research suggests that the economic vitality of the empire did not collapse as rapidly in the 3rd century as was once thought.
If we take the largest estimate of the size of the Roman army and take a pessimistic view of Roman population in 300 estimating it to be 50 million (noting that is a pure guesstimate and not based on any definite evidence), we obtain an estimate that the Roman army made up 1.3% of the population. If we employ John Lydos’s numbers we obtain an estimate of 0.87% of the population.
These numbers do not suggest that the Roman army was especially large or burdensome in comparison to other advanced preindustrial societies. In the late seventeenth century, the armies of Louis XIV represented as much as 2% of the French population of 20 million. The Dutch Republic in the 1750s also employed 2% of its population in its armed forces (45,000 out of a population of 2.25 million). The Prussian state employed around 3.5% of its population in the army in the mid-eighteenth centuries. Even Britain employed around 1.2% of its population in its army as perhaps as much as 3.8% of its population in the navy at the height of the Napoleonic wars (approximately 400,000 out of a population of 10.5 million).
What about the size of the Roman bureaucracy? It is accepted that the bureaucracy of the principate was tiny (perhaps 10,000 individuals, many of them freemen and slaves of the imperial household). The late Roman bureaucracy was substantially larger. But even if the bureaucracy after Diocletian was ten or fifteen times larger than that of the Augustinian empire, it would not meaningfully change our comparisons. At most around 1.6 % of the population would have been state employees (either soldiers or bureaucrats).
These numbers are not trivial. They certainly attest to the tax-raising powers of the Roman state. The successor states would not be able to maintain professional armies or bureaucrats at all. Nevertheless, it seems implausible to describe a state that employed less than 2% of the population as a “command economy”.
