How did agricultural productivity change in Italy with the fall of the Roman Empire and through the early Middle Ages?

tl;dr Productivity did not change much in late antique Italy. Small decreases through disruptions occurred locally and repeatedly, but small increases in techniques and efficiency can also be observed. Lamentations about the decline – of agriculture in general – match the population contraction but not the productivity of the remaining farmers. Agri deserti dominate the sources, but newly developed land has to be calculated against those often marginal areas lost. Despite our name for the regions, 'Italy' is not uniform, and Northern Italy was affected much more and earlier than Southern Italy.

The relative productivity of a plot of land in the time of transition from late antiquity to the early medieval did not change drastically due to politics or social changes. That means that total output or gross domestic product are irrelevant for answering the question, and only become relevant when explaining an answer.

"Decline in agriculture" is not congruent with decreased productivity. Neither is "developing feudalism" a driver for innovation in this area or time. The people didn't change over night, the tools and techniques did not change over night, the crops and animals didn't change over night. While production shifted places and long distance trade decreased to a certain extent.

But as some places were deserted and others re-populated, the level of technology made the level of exploitation of the land almost equally well if the whole peninsula is taken into view.

The following quote represents common, and outdated knowledge, largely superseded, by studies like those that come below:

It is generally agreed that there was a decline in agriculture in the later Roman empire, but little attempt has been made to estimate how serious it was, and on its causes debate has been inconclusive, whether it was due to the general exhaustion of the soil, to shortage of agricultural manpower, or, as contemporaries believed, partly to barbarian invasions and depredations but predominantly to over-taxation.
_{A.H.M. Jones: "The Later Roman Empire 284–602. A Social Economic and Administrative Survey", Blackwell, Oxford, 1964.}

We probably have a distorted picture of that area and time. The crisis of the Western Empire and its ultimate military downfall dominate the picture. If we look more at the base of societies, in this case agriculture, we see a much different thing emerge: the disappearance of the formal imperial bureaucracy was not that important as the history of big men and powerful warmongers suggests. Even some of the old senatorial families stayed at the top of society and organised their riches in parallel with new elites built from the officers of invading armies. The countryside was not completely ravaged, people were not just killed away nor were all connections broken down.

Life went on, agriculture continued, commerce went on, built on the foundations already established by the late Empire. The most visible and radical change was not the result of imperial defeat, but from a much more destructive force: the establishment of a monetarised system of payments. Gold and (land) speculation inherited from the time of the late Empire enabled a fierce stratification that led to a general boom in the East and South, but had a much more diversified effect in the somewhat more chaotic West, such as Spain and Italy.

We do not see constant and uninterrupted chaos, but constantly expanding commodification, monetarisation, profit-seeking, and resulting rivalries. Consequently, as estates expand and grow, others shrink or are abandoned, if they no longer make a profit. Subsistence economy retreats, wage labour evolves further.

Compared to the above factors, things like warfare, environmental problems like soil depletion or climate change are not unimportant factors, but much less so than commonly thought. The above might also look like bad bad capitalism, and in fact it is of course an important step into that general direction, but it is merely to emphasise that the outflow of for example Max Weber's dogma of late antique isolationism is a raster with to much space to fill in between.

…the resilience of the late antique rural areas is a fascinating and undeniable feature of large parts of the Mediterranean world, and one obvious implication of this boom is that the demographic pattern was more complex and differentiated than any sweeping notions of a generalized decline. It is now likely that for most of late antiquity population was on an upward climb, and that the dominant agrarian classes were able to draw on a “surplus” rural population, much as the English estates would do in the thirteenth century.

But this general trend needs qualifications, especially regarding Italy:

The ‘explosive’ growth of settlement in the rural hinterlands of the East Mediterranean9 was certainly not matched by any analogous process in either Italy or Spain. Some of the Italian evidence, for example, suggests a clear pattern of the restructuring of rural sites in the late antique period, with a reduction in site densities, accompanied, frequently, by the emergence of substantially larger establishments.

The Italian story is one of demographic contraction, though Wickham notes, ‘Every region, every city, had its own cycles of depression and revival, independent of the political and economic crisis of the western empire as a whole.’

A more prosperous countryside was not necessarily one with less inequality. Indeed, all the evidence suggests that the rural society of the sixth century was more, not less, differentiated than that, say, of the third. In some areas the peasantry was clearly deeply stratified, with a handful of well-to-do households, village leaders, distinguished from a conspicuously larger mass of landless tenants and labourers.

Commutation had a major impact on the rhythms of monetary expansion, suggesting that there were two important processes running hand in hand. As taxes were commuted to cash, the countryside was more deeply integrated into commercial exchanges. […]
This process was fairly advanced by the sixth century, and would have represented a major transformation of the economy. On the other hand, much of the pressure for commutation stemmed from powerful groups within the bureaucracy, for whom the payment of salaries in gold had the crucial advantage of yielding ‘liquid assets which could be used for all kinds of investment and especially speculation in land’. The papyri contain numerous examples of members of the military and bureaucratic hierarchies buying into the assets of existing land-controllers and themselves coming to control substantial amounts of land at various times in the later fourth and fifth centuries. In addition to which, of course, there was widespread and systematic bureaucratic profiteering, involving the kinds of speculation and rent-seeking that were repeatedly denounced in contemporary sources.

As large sums of gold were accumulated through these various mechanisms in a late Roman equivalent of ‘primitive accumulation’ that appeared to contemporaries as the unbridled dominance of public officials, a new aristocracy emerged out of the expanded governing class of the fourth century, no longer merely an ‘aristocracy of office’, though it was always that as well, but an economically powerful and socially dominant group of businesslike landowners who dominated their respective regions. Unlike the west, where a small élite of the old senatorial families was hemmed in by deepening fragmentation16 or displaced by the advance of a new local ‘landed aristocracy of military officers’,17 the eastern provinces knew no juxtaposition of this sort, and the emergent aristocracies of (mainly) the fifth century were both more recent than the ‘old’ senatorial families of the western empire, and more precocious than the landholding military officers of sixth-century Italy. Indeed, it is difficult to speak of a ‘crisis of the old ruling class’ in the case of the East Mediterranean, where the native municipal élites were simply no match for the powerful new elements who invaded the landholding structure and became the ruling class. In short, the late imperial state nurtured a new ruling class, much as the Stalinist regimes sought to do in parts of Europe in the twentieth century (but there unsuccessfully), and this process is particularly evident in the eastern provinces.

_{Source of quotes: Jairus Banaji: "Agrarian Change in Late Antiquity Gold, Labour, and Aristocratic Dominance", Oxford University Press: Oxford, New York, 2001.}

Between 489 and 493 further warfare hit Italy, resulting in the Ostrogothic tribe becoming masters of the peninsula. Little hope, one would assume, for the wretched province of Italy. And yet the Ostrogothic takeover, far from seeing the peninsula slipping into barbarous ways, in fact marked a cultural pick-up, with king Theoderic’s strong and long rule stabilizing Italy’s position on the economic market, and with various marriage alliances to Germanic neighbours blocking the likelihood of military conflict. With towns back in business, a good degree of building activity is registered — notably in those frequented by the royal family, notably Verona, Pavia, and Ravenna — and it coincides with a major upturn in rural productivity (Wolfram 1988, 284–9; Lusuardi Siena 1984).

Our principal source for this period, covering c.495–535, is the lengthy but invaluable collection of letters by Theoderic’s chief minister, Cassiodorus Senator. His ancestral home was in southern Italy, and it is no surprise to find that a good deal of information on this previously much-neglected region suddenly emerges. Cassiodorus proudly highlights the excellent stocks of horse and cattle in the hills of Bruttium, the healthy supply of Lucanian pork to Rome, and the increasing popularity (thanks to his advertising campaign) of Bruttian wine (Variae, 8. 31, 33; 11. 39; 12. 12). His tourist board act suggests a happy revitalization of the town- country relationship, in southern Italy at least; for the Squillace district, besides a fine climate, he says that

residents in the city are not deprived of the fine sight of workers in the field. They look out to their satisfaction on abundant grape harvests; on the threshing-floors, productive work is in their view; the olives, too, display their greenery. No one lacks the pleasures of the countryside who can see all this from the city. (Var. 12. 15)

His comments, in fact, seem to be borne out by villa excavations in the south, notably those at San Giovanni di Ruoti, where major economic prosperity is evident in the late fifth century, extending into the second quarter of the sixth (Small 1991). In Samnium the small villa located beneath the eighth-century monastery at San Vincenzo al Volturno also relates to this period (Hodges 1988)— though admittedly other Roman sites in Samnium, such as San Giacomo or Matrice, show no sign of partaking in this relative boom.

A good indication of rural health is witnessed by Cassiodorus’ documentation relating to renewed pressure on arable land. For example, a letter of c.510 records the patrician Decius wanting to drain part of the Pontine marshes at Decemnovium, nineteen miles north of Terracina, featuring a stretch of the Via Appia. Decius’ workmen have a hard task:

the marsh ravages...the neighbourhood like an enemy.... It is a notorious desolation of the age, which, through long neglect, has formed a kind of marshy sea, and, spreading by its waters a hostile deluge over cultivated land, has destroyed the kindly arable equally with shaggy woodland. Since it began to be exposed to the marshes, the soil has been robbed of its crops, and nourishes nothing useful beneath the water. (Var. 2. 32; cf. 2. 21)

In discussing the late Roman empire one all too readily thinks of a rapid collapse that overwhelmed all and was perceived as such by all. But this was no modern-day Soviet Union or Yugoslavia with news networks allowing for the instantaneous transmission of events, ideas, and fears.

But towns continued to exist, politicians and clergy still wrangled, and armies still guarded frontiers. To feed them, of course, land needed to be farmed. The documentary sources, when explicit, tend to paint a bleak picture of the landscape— barren, cold, and lost to nature—but in land grants, land disputes, and church and monastery foundations enough scattered hints emerge of maintained activity in the countryside involving villages, farms, estates, and shepherds. What modern commentators tend to do is to underestimate the resilience of the peasant, or rather of country-based folk: their livelihood had always lain in the land, and it is wrong to assume an overriding readiness on their part to move up permanently to distant, often inhospitable heights.

_{Neil Christie: "Barren Fields? Landscapes and settlements in late Roman and post-Roman Italy", in: Graham Shipley & John Salmon (Eds): "Human Landscapes in Classical Antiquity. Environment and Culture", Routledge: London, New York, 1996. (online)}

Continuity of commercial wine production in many regions of Italy and continuing trade in Italian wines both inside and outside Italy is clear up to the Lombard invasions. Technological continuity with the earlier Roman period is strong, with horizontal lever presses, using stone weights, remaining in common use. There is little evidence for technological innovation during this period.

In the Late Roman period, much of the evidence points to commercial production based on the great villas of aristocratic families. The transformation and abandonment of some villa sites in the 5th and 6th centuries, brought about by fundamental social and political change, must have caused some disruption to these centuries-old patterns of agricultural production and commerce. Yet, as Lancon notes, even tumultuous events like the Gothic and Vandal sieges of Rome in the 5th c. and the Byzantine wars of the 6th c., may not have seriously affected the traditional exploitation of the Italian countryside. The pattern of wine production hinted at in the pages of Symmachus is still visible in the writings of Cassiodorus, whose family’s vineyards in Bruttium, now part of a newly-formed monastic community, continued to produce wine on a commercial scale in the early 6th c. By the end of the 6th c., the organised trade in wine may have diminished, but Italian vineyards continued to produce wine for local markets using technologies little changed from the time of Pliny.

_{Jeremy Rossiter: "Wine-Making after Pliny: Viticulture and Farming Technology in Late Antique Italy" in: L. Lavan, E. Zanini, and A. Sarantis (edd.) Technology in Transition A.D. 300–650 (Late Antique Archaeology 4 – 2006) (Leiden 2007), pp. 93–118.}

In reality, the tendency is evident to value and praise the productive systems of each territory (grain and wool for Apulia, pork for Lucania et Bruttii, wine for Bruttii), in an area that enjoyed a long period of relative tranquillity and security during the crisis of the 3rd century and also during the following century. One touches on an aspect of extreme importance that clearly distinguishes the southern regions from those of the center-north, which experienced, on the contrary, a destructuring of the economy and rural settlement already by the 2nd and 3rd centuries. This is an element that, associated with its fortuitous central position in the Mediterranean and its good connections on land and sea, constituted a strongly attractive factor for investment on the part of the rich Roman senatorial aristocracy and also local elites in the 4th and even more strongly in the 5th century, as demonstrated by the dense rural population in various areas of Southern Italy.
While elsewhere in Italy the process of destructuring inexorably rose during the 5th and the 6th century, the Southern Italian provinces were indeed the last enclave of the aristocratic property and the late roman economic development tied to agriculture, to sheep-rearing, to crafts and trade.
One of the first pieces of data offered by systematic archaeological survey projects concerns the quantitative evolution of rural settlement between the early and middle Imperial period and Late Antiquity. From this evidence it is difficult to distinguish specifically the 5th century from the general period of Late Antiquity, which encompasses also the 4th and the first half of the 6th century.
In the study of the late antique countryside in Italy, the generalized rarefaction of villas, and of rural settlement in general, has been confirmed for this period when compared to the first centuries of the empire. Tamara Lewit has estimated the median levels of abandonment in Italy at 67% already by the second half of the 4th century. However, the reduction of sites could be made up for by the expansion of rural structures, part of the general process of the concentration of property. Abandonment, then, favoured some villas, which became larger and more luxurious: this fact has been confirmed absolutely.

_{Giuliano Volpe & Maria Turchiano: "The last enclave. Rural settlement in the 5th Century in Southern Italy: the case of Apulia", 2007.}

The study of faunal remains is contributing the most to our understanding of the role of animal husbandry in the late antique economy, and its change in the period of transition between Roman and Post Roman economies. In the West, the 5th and 6th c. seem to have brought about a shift away from specialised animal rearing, towards more mixed husbandry practices. Pollen studies have confirmed a similar trend toward mixed farming: cereal production receded in most marginal regions, where it was replaced by reforestation and a return to seasonal pastoralism. Rather than demographic decline, the regression of cereal production and a return to mixed farming were likely due to the demise of the imperial demand for agricultural goods, and the end of an integrated market that made it possible to profit from crop specialisation.

_{Andrea Zerbini: "The Late Antique Economy: Primary and Secondary Production", in: Luke Lavan: "Local Economies? Production and Exchange of Inland Regions in Late Antiquity", Brill: Leiden, Boston, 2013.}