16

Economy and supply

Roman economies

We have seen how London was brought into existence to support the Roman conquest of Britain, and that the logistical supply of Rome’s armed forces remained a critical concern during a succession of campaigns aimed at the pacification of Britain. Even when imperial attention was distracted by the Dacian wars at the end of this period of advance, and the British garrison reduced, the loyalty of the remaining troops was retained through investment in their welfare and supply.

Despite the opportunities for wealth creation carried by the flow of resources through London, the early Roman city boasted few monuments of elite patronage. This suggests that the civic euergetism of a land-owning aristocracy remained unimportant to political life. London was unusual in this regard, making it a poor fit for interpretative models based on Mediterranean city politics where the exercise of Roman power was mediated by local elite society. In Chapter 3 we noted how Moses Finley’s studies of the ancient economy encouraged a minimalist view of the role of the Roman state in civic affairs, contributing to a scholarly reluctance to credit Roman authorities with a determining role in the foundation of London. This was the outcome of a long-standing debate over the nature of the ancient economy, where opinion divided over whether development can be understood as the consequence of the working of market forces, or whether such forces were fundamentally constrained by social and political factors. ¹ A key concept within this debate is that of the ‘consumer city’: an ideal type that characterizes cities as places where aristocrats enjoyed a political life sustained by rents and taxes drawn from rural estates, and where trade and industry remained subordinate to their interests. ² Although still enormously influential, it is now recognized that Finley’s model underestimated the scale and complexity of ancient trade, overstated the uniformity of the ancient world across time and space, and failed to acknowledge the extent to which imperial expansion promoted economic growth. ³ More recent studies have turned to ideas developed in the field of New Institutional Economics in order to explore the impact of different institutions of the Roman state on economic performance. ⁴ These new avenues of research have stimulated a concerted attempt to characterize and quantify economic growth in the Roman Empire. ⁵ The implications of this work for the study of Roman London are considerable, offering analytical tools that encourage us to treat London’s early urban development as a product of imperial strategy.

Military demand

Rome’s military success stemmed from its ability to support large concentrations of professional troops, establishing supply arrangements that also underwrote the needs of large cities. Whilst the importance of military demand to the economic development of Rome’s frontier provinces is easily recognized, it is harder to pin down the detail of the mechanisms employed. ⁶ The respective contributions of market provisioning and directed supply are much discussed: were goods moved because traders found profit in meeting the needs of cash-rich army communities, or as official rations secured from tax and rents raised in kind that could be supplied to troops without consumer purchase? ⁷ Both forms of supply are attested in our fragmentary sources, and since the logistical needs of the armed forces were met on an ad hoc basis involving much improvisation, different situations would have encouraged different solutions. In a sense it is not critical to our understanding of London’s economy to know how much supply was the consequence of the directed ‘push’ of rations issued by the state, and how much the aggregate ‘pull’ of purchasing decisions made by army units and individual soldiers. In either case, London’s port was the critical pumping mechanism, speeding supplies along a network of arterial haulage routes.

In order to assess the importance of London’s role in military supply, it helps to briefly review what is known about the scale and character of military demand. A campaigning army had immediate call on a baggage-train of supplies known as impedimenta, with longer-term needs met by the organized shipments known as commeatus. ⁸ Soldiers and their mounts required equipment, food, fodder, and firewood. Whilst local supplies were requisitioned and foraged, major campaigns relied on amassing significant reserves in advance of troop deployment. The staple was grain, delivered to troops as unground wheat, bread, or biscuit. This was accompanied by a meat ration that might consist of salted, smoked, or fresh veal, mutton or pork. ⁹ Soldiers additionally consumed copious quantities of wine (usually as sour wine or vinegar), olive oil and salt. There was considerable variation over which of these foodstuffs formed part of the regular dole, the costs of which were deducted from pay, and which the soldiers had to purchase on their own behalf. At one point the daily ration issued to soldiers in Egypt was described as consisting of two pounds of bread, a pound and a half of meat, a quart of wine, and half a cup of oil. ¹⁰ Late second-century ostraka from the same province recorded cash payments received by soldiers in lieu of the wine that was their due from official rations. ¹¹ In addition to routine supplies, soldiers and officials also benefitted from special distributions of free meat and wine made at religious festivals and sacrifices. ¹² The scale and frequency of such distributions should not be underestimated.

Writing tablets found at Vindolanda on Hadrian’s Wall describe various cash-based transactions that secured supplies for the military community there between c. ad 85 and c. ad 130. ¹³ Threshed grain and hides were purchased locally by middlemen, reducing reliance on long-distance supply. ¹⁴ At least one of the dealings involved Caesariani, the imperial freedmen and slaves who managed imperial assets and had command of goods raised as tax and rent in kind. ¹⁵ These documents were written to formalize business agreements, and perhaps biased towards recording market transactions, but leave no doubt that civilian traders played an important part in the fort’s supply. These traders were not serving soldiers but remained part of a wider military community. Commanding officers were responsible for the wellbeing of the troops under their command, and may have used payroll budgets to purchase goods from contractors who may, in turn, have acquired some supplies from imperial sources.

Drawing on what we know of military diet, James Jones calculated that the earlier campaigns in Wales, involving about 20,000 men, needed the equivalent of 900 shiploads a year of grain, fodder, wine, and oil. ¹⁶ Prudent planning involved building sufficient reserves to cover the needs of the campaign season. Some provisions may have been found locally, and coastal supply was developed to avoid trekking goods through the Midlands, but in the early stages of advance much of this material is likely to have arrived via London since this was the hub where coastal routes intersected with the inland road network. The stores of imported grain in the Neronian city (above p. 74), may been destined for overland carriage along Watling Street.

These supplies were an important concern of the provincial governor, but the direct responsibility of the procurator. Their issue anticipated the later Roman institutions of the annona militaris, used to organize the delivery of the soldier’s rations and drawing on goods raised as tax in kind. We don’t know the extent to which these later arrangements developed from solutions instituted to guarantee supplies to active forces in earlier phases of Roman imperial expansion. What we do know is that Rome developed a bureaucracy to police the contracts that kept its forces fed and equipped. ¹⁷ The institutional frameworks and physical infrastructure that this engendered were available for the supply of other goods and other consumers within the orbit of the provincial government, such that official rations were sometimes issued to civilians and slaves including those working on major construction projects. ¹⁸

The first concern of the Roman state was to satisfy the needs of its legions, guaranteeing food security and sealing their loyalty to the emperor. Military campaigns placed an enormous burden on supply mechanisms at times when manpower was in greatest demand. The higher command relied on contracted middlemen (navicularii, negotiatores, and mercatores) to move goods to their destination. ¹⁹ These merchants and shippers added capacity without depleting the forces at the governor’s disposal. The entrepreneurial activities of this business community were underwritten by the contracts and facilities afforded by the patronage of military commanders and administrators. Long distance trade was shaped by these official supply routes and patronage networks, even when it involved privately owned cargoes destined to be sold for profit on the open market.

Wherever the army went, merchants and contractors followed. Tombstones and inscriptions suggests that in the north-west provinces businessmen (negotiatores) formed a powerful community. Lyon was an earlier hub for their activities which stretched through Gaul, Germany, and into Britain along the Rhone-Rhine axis. Votive inscriptions witness the presence of Negotiatores Britanniciani and specialist traders in particular products along this route. Koenraad Verboven has described the emergence of a business class of merchants, producers and traders that relied on the market represented by the military garrisons of the north-west frontier. ²⁰ These private operators were part of a community in which distinctions between military and civilian could be highly blurred. ²¹ Veterans and freedmen were perfectly placed to benefit from military contracts and some soldiers were involved in private business ventures whilst in military service. ²² Shippers (navicularii) could be hired to carry state-owned annona cargoes. Navicularii can be distinguished from negotiatores or mercatores in that their duties did not require them to take ownership of the goods they carried. They could work individually or within consortia, using their own or chartered vessels. ²³ Shippers were able to take advantage of capacity developed around the needs of particular campaigns or building programmes, as well as for the seasonal haulage of agricultural surplus raised as tax or rent in kind. The logistical infrastructure on which they relied was in large part army engineered. Additional goods could be carried piggyback with bulk shipments and as makeweights to fill empty space on transports returning from officially commissioned deliveries. ²⁴ These circumstances, allied to the considerable purchasing power of soldiers and officials, contributed to unprecedented levels of trade.

The preparations for the campaigns in Wales described by Jones would have involved docking and unloading an average of 7.5 ships every day during the 120 sailing days of the season (below p. 204). The routine provisioning of London would have been an equal task, with the urban population rising to as many as 30,000 by the middle of the second century. ²⁵ Supplies of building materials for public construction projects and fuel for the public baths added significantly to river traffic. On busy days, dozens of ships were carried into London by the tides, with balancing numbers making their exit. Most cargoes served official needs, but the attendant infrastructure and institutions opened the way for a host of other trades.

Elsewhere, the most significant example of official intervention in underwriting food security were the yearly imports of grain to Rome known as the annona. ²⁶ This scheme shows how Rome approached the issue of administered supply. Although originally used as a reserve to protect Rome from shortages, the annona came to involve free monthly distributions of grain to the heads of urban households through the frumentationes. There was a tendency to centralize and extend the annona through time. ²⁷ At some point large quantities of olive oil were imported into Rome in similar fashion, eventually being added to the dole under Septimius Severus. Imports of Spanish olive oil contributed to a massive dump of broken amphorae at Monte Testaccio in Rome, where control marks suggest that the imperial authorities purchased olive oil from producers in order to guarantee its supply to Rome. ²⁸ Until distributed free under Severus, this oil may have been sold at a fixed or subsidized price or simply released onto the market at times of shortage.

Rome’s annona was administered on the emperor’s behalf by a dedicated public official (the praefectus annonae), combining produce taken as rent or tax in kind with market purchases. ²⁹ Shipments were made by private merchants, who gained tax privileges in return for their involvement. ³⁰ Massive harbour facilities were built at Portus to accommodate the annona, and we have seen that similar arrangements might account for some phases of building activity at London’s port (p. 101). Although not a commonplace measure, cities other than Rome stockpiled grain with the support of the imperial government, giving the authorities the means to stabilize prices by releasing reserves onto the market at times of scarcity. ³¹ London’s vital role in administered military supply, frontier vulnerabilities, and close association with the provincial government, may have combined to occasion such intervention. This might also provide a context for late first-century investment in buildings associated with food supply and distribution within the city (p. 151). Changing patterns of supply in later Roman Britain are also best explained in the context of third-century reforms to a provincial annona designed around the needs of the army and government (p. 327).

Long-distance supply

London was a hub for the trans-shipment of products imported to Britain across the English Channel. This is seen in London’s pottery assemblages, where imported goods included amphorae, Samian, other Gallic and Rhenish fine wares, and mortaria. This pottery is our main source of evidence for long-distance supply, but a poor proxy for the larger volumes of perishable materials that must have been transported. Nearly a quarter of the pottery found in first-century deposits in London was imported across the Channel, compared to 10 per cent or less in most other Romano-British towns. ³² Amphorae used to carry wine and oil produced in Gaul, Italy, and the Mediterranean, as wells as concentrated grape syrup (defrutum) and fermented fish sauces (garum and liquamen), make up about 40 per cent (by weight) of all Neronian pottery in London. They are not evenly distributed across the city, but concentrate in the port and forum where they are almost twice as abundant. ³³ Fewer amphorae reached the outer parts of the early town, suggesting that their importation wasn’t primarily aimed at supplying these urban communities.

The commonest type of amphorae found in London, the Dressel 20, contained olive oil imported from Baetica in southern Spain. Studies of its distribution show that this reached Britain following military supply routes along the rivers Rhone and Rhine. ³⁴ José Remesal Rodriguez has argued that the oil was acquired through tax and rent in kind, transported by shippers employed by the state and supplied to troops in exchange for sums docked from pay. He supposes the involvement of the praefectura annonae, responsible for the grain supply of Rome, in managing the exercise. This has been questioned in more recent studies where the importance of market-based supply has been reasserted. ³⁵ These arguments over the mechanisms involved, important as they are, don’t change the fact that the bulk transportation of olive oil was directed to places where the legions were deployed, shadowing other officially sanctioned supplies.

A significant volume of wine also reached London in large wooden barrels sealed with pine resin. Examples made of silver fir, probably imported from the Rhineland, were reused to line wells and others were broken up and trimmed to make writing tablets. ³⁶ Many were stamped with the names of shipper, producer, or retailer and an example from One Poultry was stamped across a bung to prevent tampering. ³⁷ Army-influenced tastes are likely to account for other particularities in imports to London, and involved extending supply routes from the Rhine to the Thames. Examples include the clay lamps that are more common in early Roman London and Colchester than elsewhere in Britain and the quern- and millstones imported from Germany. Various other luxuries were imported against the needs of wealthy residents, perhaps sometimes as private property accompanying high-status immigrants and otherwise as merchandise. These included textiles, ivory bracelets, amber beads from the Baltic, and gold and emerald necklaces. A variety of new and exotic foods were also imported. ³⁸ A shop destroyed in the Boudican revolt sold imported spices, herbs and fruits (above p. 80). Elsewhere black pepper, an import from across the Indian Ocean, pomegranate, peaches, olives, figs, grapes, cucumber, edible stone pine kernels, and walnuts have been identified.

Ragstone, flint, and chalk were imported from Kent for use in concrete foundations but were unsuited for most stoneworking. Kevin Hayward has reported on the wide variety of building stone imported to London. ³⁹ This included Reigate stone and Hassock greensand, and the dark brown-black ‘Purbeck Marble’ from Dorset. High-quality oolitic limestone from the Cotswolds, Cambridgeshire, Lincolnshire, and Northamptonshire provided the detailed carvings on major monuments. London also imported stone from quarries on the French coast. Clay roofing tiles and bricks came mainly from Hertfordshire and local brickfields, and some Yorkshire roofing slates have also been found.

Relatively large quantities of tablewares were imported into Britain through London. Martin Pitts describes a series of such finds commonly associated with military and colonial establishments including Lyon ware beakers and cups, decorated terra sigillata, Gallo-Belgic bowls, glass vessels, and mortaria. ⁴⁰ The most ubiquitous of these was the glossy red Gallo-Roman sigillata, or Samian. London’s late first-century pottery assemblages contain unusually high percentages of Samian, sometimes forming over 10 per cent by weight. ⁴¹ Lyon ware and terra sigillata were particularly popular with communities associated with the military and colonial infrastructure along the Rhine, and which reached into Britain after the conquest. Distribution patterns explored by Nico Roymans suggest that the soldiers stationed at the earliest Rhine military bases probably owned a cup and platter of terra sigillata for individual use, with larger sets used by officers when throwing banquets. ⁴²

At various points in time, London was the major port of entry for Samian. During the second century large quantities of these glossy red platters, cups and bowls were stored in warehouses next to the quays where they had been landed. London may have been the most important single destination for products of the Central Gaulish workshops of Les Martres-de-Veyre, with only small quantities ‘leaking’ out from supply route along the Rhine. Mike Fulford has considered how Samian could reach Britain without any tail-off in volumes of supply. ⁴³ Where market supply prevails, the costs of transportation are added to the price charged to consumers and will depress sales at more distant locations. Fulford suggests that transport costs may have been subsidized or met in full by the state when the vessels were included with equipment issued to senior soldiers and government officials, the costs of which were deducted from pay. He also suggests that the Samian that travelled beyond London may have been transported along official supply networks set up through the posting stations of the cursus publicus, which offered points from which these goods reached more remote consumers. Allard Mees suggests that the aggregate buying power of regular soldier’s pay offers a less complicated explanation for the bulk movements of Samian, questioning Fulford’s ‘subsidized distribution’ argument. ⁴⁴ He is heavily influenced, however, by the idea that London’s import of Samian witnesses the importance of civilian markets. He suggests that onward supply from London was the consequence of city merchants selling goods to customers who came to the city. This is not the conclusion drawn here. The dendrochronological evidence shows that investment in London’s port and transportation infrastructure coincided with preparations for military campaigns and phases of annona reform aimed at retaining the loyalty of the Roman forces in Britain. Some of these episodes of official investment related to changes in the pattern of long-distance supply routed through London, which impacted on the source and volume of the Samian brought to the city.

This doesn’t require us to believe that administered supply was the dominant mechanism. It is possible to imagine circumstances in which new colonial and administrative settlements, major public building programmes and particular festivities and celebrations were accompanied by subsidized supplies arranged by the administrative authorities. These could have been sponsored by the governor or procurator, involving the use of public slaves and troops, and perhaps state-transportation at times of surplus capacity. At other times, particularly those of scarcity or untoward need, public authorities may have commissioned private contractors to arrange additional supplies, sometimes moving imperially owned goods (as shippers) and sometimes bringing to market the goods that they had acquired for this purpose (as merchants). Between these officially subsidized distributions, local arrangements and market supply is likely to have prevailed. This diversity of approach absolves us of the need to fully endorse or reject the models that have been brought to bear on the tortured issue of how state control intersected with market forces.

Monetary transactions

As noted in Chapter 7, Neronian Samian appears to have been sold from a shop at One Poultry alongside other exotica used at dinner parties, including spoons and spices. The demand for these table luxuries stemmed from patterns of consumption that also relied on imported oil and wine. Martin Pitts has described how these elite tastes had derived from military communities along the Rhine. ⁴⁵ They don’t necessarily identify military consumers, but were found in the wider patronage networks attracted to command sites in general and London in particular. City shopkeepers met some of this demand, presumably through the offer of fairly priced goods. Shops and the emergence of a monetary economy, supported by state intervention to facilitate the movement of vital supplies, all contributed to the expansion of market activities and stimulated growth. ⁴⁶

London’s commercial transactions relied on coin. It has been estimated that some 6.5 million denarii needed to be imported to Britain each year to pay the soldiers and officials in imperial service. ⁴⁷ These were followed into currency by smaller bronze denominations that facilitated low value transactions within the military and administrative community. The cumulative spending power of employees of the Roman state was the gateway through which currency entered the economy. ⁴⁸ Coin use and coin loss was consequently higher in towns and at military sites, where it seems reasonably certain that the economy was fully monetized. Urban growth depended on the transactions facilitated by the pervasive use of coin, set in train by the presence of troops and officials with cash to spend and commissions to meet. Elsewhere the spread of coinage was limited, and supplies of small change were insufficient to have supported their routine use. In any case the face-value of the available coin was skewed way above levels useful for the rural poor. ⁴⁹ Early Roman Britain can consequently be described as presenting a landscape of ‘coin using islands…in an overwhelming sea of virtually coinless peasants’. ⁵⁰

We are reminded of the importance of small change to London’s daily life by a record of the payments due to a haulier described on one of the writing tablets found at the Bloomberg site. ⁵¹ This text, dated 21 October ad 62, which we have already considered in the context of post-Boudican rebuilding (p. 90), recorded a contract made between Marcus Rennius Venustus and Gaius Valerius Proculus to bring twenty loads of provisions to London from Verulamium by 13 November. Roger Tomlin suggests that this timetable might represent the operation of a single waggon going to and fro between the cities on a daily basis, but this could not have been achieved within the allotted time using ox-drawn transport given the slow speed of such transport. It seems more likely that two waggons were employed, one travelling in each direction: the distance of about 30 kilometres could just about be covered in the ten hours of daylight available at that time of year. The contract agreed a transport charge of one quarter denarius, or four asses, for each shipment. This would result in a day-rate of two asses: a modest sum, given that a soldier’s rate of pay amounted to ten asses a day, indicating that the haulage contractor had access to low-cost labour. Roger Tomlin suggests, from his reconstruction of conditions attached to the contract, that three asses were paid on receipt of each transport, with the remainder held as a retention against completion. It is difficult to see how this arrangement could have worked without coin changing hands on the safe receipt of each consignment, presumably at the forum or city gates. This reliance on small-change to pay contracted hauliers, and lubricate a host of similar routine transaction, might explain why it was sometimes necessary to manufacture low-value irregular coin (pp. 79 and 330). The value of this small-change was amplified by the access it provided to the carriage goods.

The use of coin also simplified the collection of the port duties levied on imported goods. ⁵² Tax income from the transportation of routine goods was considerable, as shown by customs-house registers from Egypt. ⁵³ It is therefore possible that investment in transport infrastructure in London was not only aimed at easing bulk haulage associated with the annona, but at improving tax collection on goods entering the city. The riverside quays controlled access to and from the river. A parallel can be drawn with restrictions on river traffic imposed under Queen Elizabeth I, who limited the unloading of goods to a number of legal quays where customs duties could be collected. ⁵⁴ Public investment in the construction of docks could become profitable where this gave control over lucrative port duties, just as the adjacent warehousing was a source of rental income. Various transactions were taxed, generating income for the state. Inscriptions from neighbouring provinces suggest that efforts to improve on tax collection were underway in the late second and early third centuries. We know that such taxes were levied at London late in the Roman period, since lead seals used to identify tax exempt imperial property have been found on the Thames foreshore (below p. 301). ⁵⁵ Merchants carrying state goods did not have to pay these customs (portoria), and goods belonging to veterans and military personnel could also be exempt. ⁵⁶ Shippers are likely to have taken advantage of these tax waivers to import products alongside official supply cargoes. ⁵⁷

London and the Roman economy

Gustav Milne has suggested that many of the luxuries imported into early Roman London were brought to meet the needs of the town’s richer inhabitants, and that the town itself was the magnet that attracted exotic materials. ⁵⁸ It is difficult to believe, however, that London’s thriving economy was sustained by internal demand alone. It is more likely to have prospered because of opportunities to profit from military contracts and subsidized links with Germany and Gaul. As a map of Roman roads clearly illustrates, London was the principal nodal point in the network of control that Rome established over southern Britain. The needs of long-distance supply justified building and maintaining an infrastructure of quays and warehouses at this hub. Rome’s earlier experiences of territorial expansion meant that such needs were well understood and could be addressed through strategic planning. London’s growth was the product of this determined state involvement in establishing the physical infrastructure and institutional frameworks conducive to trade, and which included the engineering of transport infrastructure, creation of storage capacity, and market intervention to secure the flows of staples and luxuries on which Roman communities relied. ⁵⁹ London would have been modelled on the earlier example of places such as Lyons (above p. 71), which played a pivotal role in military supply and civilian commerce where facilities were developed to support the Roman forces. ⁶⁰

Capital expenditure on public building projects added to the flow of resources into London. Urban markets helped to convert local surplus—obtained as tax, rent, or tribute—into army supplies, defraying their cost and underwriting security. London’s shops and workshops made it possible for contractors to transform other produce into goods for sale, for conversion into coin and thence into the bullion on which the state ultimately relied. The cash-using consumers found amongst this urban population created additional opportunities for economic growth. Many of the activities now focused in London’s port and town were highly labour intensive, as the next chapter describes, and stimulated population growth. The procurator stood at the heart of this rapidly expanding economy. The contractors that his patronage drew to London would have claimed a major role in collecting, storing, and shipping both annonary goods and the ancillary luxuries that facilitated competitive social display amongst the ruling elite. The annona supplies, and a complex range of attendant institutional arrangements, would have both powered and distorted London’s Roman economy. ⁶¹

David Mattingly has described how Rome’s market economy operated alongside socially constrained and politically dominated systems of exchange. ⁶² The military supply of the frontier provinces formed part of the political economy, which also relied on exploiting mineral and resources obtained from extensive imperial and public estates. The logistical demands of exploiting and transporting these resources required the state’s involvement in a redistributive exchange system that defied normal economic logic. On the margins of this directed economy, lay a separate world of subsistence and peasant farming that made limited use of the cash economy or of the goods available through urban markets.

London changed as Roman power became established, and it is possible to explain phases in the evolution of the port as a product of the institutionalization of annona supply. The conquest phase was characterized by a relatively low investment in port infrastructure, despite the importance of London in the business of supply. The military reoccupation of London after the Boudican revolt prompted the construction of new harbour facilities that were then episodically extended and improved down to the middle of the third century. Some phases of port construction may have involved direct military support, drawing on the engineering capacity of auxiliary forces commanded by the governor.

Reforms of the early Flavian period appear to have been aimed at improving government control over the extraction, movement and management of resources obtained regionally, reducing dependency on long distance supply and establishing an infrastructure aimed at a making the rule of Britain more sustainable. Such reforms are likely to have involved tightening-up the administrative apparatus used to raise tax and rent, and ensure its speedy delivery to military consumers through the transport infrastructure. The development of London in this period may, in part, have been at the expense of the supply-base established at Richborough on the south coast. It has been suggested that the military use of this site was reduced from as early as c. ad 70 before ending c. ad 85. ⁶³ These Flavian reforms penetrated deep into the countryside in south-east Britain, which witnessed a significant change in the character and range of ceramic use at this period as Gallo-Belgic patterns of elite consumption gave way to styles of consumption and display imported to Britain from the German frontier. ⁶⁴ This may also have put in place the essentials of a market-based economy, albeit one that was largely restricted to a money-using elite community soldiers, officials, landowners, and middlemen.

London’s early development, down to c. ad 85, can best be explained through reference to military campaigns planned in Britain. After that date sporadic attention may have been given towards the infrastructure of annona supply, as implied by the waterfront building programmes of c. ad 94–8 (arguably ad 97) and c. ad 101–4. These works were undertaken at times of wider reform of Rome’s supply infrastructure, under Nerva and then Trajan. These later phases of investment in London’s harbour might be considered a form of path dependency, in which the existence of institutions aimed at supplying goods imported from the Mediterranean became an essential part of the relationship between emperor and the military forces, resulting in continued investment in port infrastructure beyond the point where there was clear economic or strategic benefit to these arrangements. The concept of path-dependency may also help explain the radical nature of third-century changes to London’s economy, when the inherited inefficiencies of an unnecessarily burdensome supply system became unsustainable in the face of exogenous stochastic shock. ⁶⁵