If the late antique economy was so robust, why did it not ride out the troubles of the seventh century?One answer would be that it was already in difficulty by the mid-sixth century. The plague which hit theMediterranean world in 541 cut the population by perhaps as much as half, and in a preindustrial economy, the number of people is the basic factor which underlies output and growth. Climate change may have been a factor, too, possibly involving a shift to a warmer and drier phase. There is also evidence from tree rings for a ‘dust-veil event’ – in other words an event sending dust into the atmosphere sufficient to reduce the amount of sunlight reaching the earth – in 536–7, caused either by a massive volcanic eruption, or a comet hitting the earth. The sources seem to suggest more frequent earthquakes and famines in these years too.Military defeat by the Persians and Arabs may therefore have simply been the factor tipping an already fragile economy into profound recession.35 These arguments attract considerable support, but equal criticism.Many familiar with the relevant evidence are not convinced that it shows an economy in trouble after 550. In Syria, Jordan, Palestine and Egypt it is easy to find evidence that suggests continuing prosperity right through the sixth century and beyond. Coinage remains common and on sites such as Pella and Gerasa in Jordan, or Caesarea Maritima and Scythopolis in Israel, the ceramic record carries on unbroken through the sixth century and through the seventh too. On the limestone massif in northern Syria it appears that the building boom had halted after 550, but dated silver treasures from the area imply no lack of resources, and the villages appear to have been occupied much as before for centuries to come. Further south, in theHawran and the Jordan valley, building-work continued. Investment in fine mosaics and new churches could be seen as a fearful response to a worsening world, but most historians would be happier to see these buildings, more prosaically, as a sign of people with resources to spare.36 The plague is equally problematic. Assuming that its mortality was as high as usually supposed, comparison with the impact of the Black Death on fourteenth- and fifteenth-century Europe would not necessarily suggest disastrous economic consequences. High mortality is likely to benefit survivors, who can take advantage of higher wages and more available land. Unless there is some other factor regulating numbers at a lower level – as seems to have happened in fifteenth-century England for example – populations will tend to bounce back to their previous level.37 The same sort of points can be made about earthquakes, famines and dust-veil events. Things happen; and life goes on. If 536–7 was a ‘year without summer’, then so too was 1816.38 Both Justinian and his critics saw plague, famine and earthquake as events of cosmic significance. It is not necessary that we should feel the same. In many ways we know more about the late antique economy than we do about that of the middle Byzantine period that followed. Yet the evidence is still too patchy and imprecise to settle these arguments, and that is particularly true for those territories of the late antique world that came to form medieval Byzantium. We know more about areas such as the Levant or Egypt, which in the seventh century were to pass under Muslim rule, than we do about Asia Minor, the Balkans, or Constantinople itself. Even the evidence from sites such as Aphrodisias, Sardis and Ephesos, which do not have modern cities on top of them and which have undergone decades of excavation, is ambiguous. On each of these sites, opinion is divided as to whether the evidence shows continuity through the sixth and into the early seventh century, or a down-turn after 550.39 And since everyone would agree that a city needs to be seen as a single economic unit with its rural hinterland, the lack of any survey looking at the hinterland of these sites is a drawback. At Sagalassos, Marc Waelkens’ Belgian team is pointing the way with an integrated programme of urban excavation and rural survey, but even here the implications are debatable. If investment in substantial new buildings seems to have stopped after 550, Sagalassos was still occupied up to the seventh. Is this evidence of a widespread ‘crisis’ in sixth-century Anatolia, or the local consequences of an earthquake in 518 that had deprived Sagalassos of a suitable water supply?40 Another approach to the late antique economy is to examine its relationship with the state. It is beyond question that in pre-modern terms the Roman empire was a powerful state, capable of having a major impact upon its citizens’ lives. What effect did the state have on the economy? Did the state’s collection of goods in kind through taxation and their transport across the empire by means of subsidised shipping – in other words the annona system (see above, p. 11) – subvert the market to such an extent that when the state collapsed, the late antique economy went with it?Or should we be looking at the economic impact of state salary payments, skewing investment and development in peculiar ways? In any case, was the apparent prosperity of the late antique world effectively a creation of the state and hence vulnerable to the political and military crisis that unfolded in the seventh century? The scale and archaeological visibility of the annona makes the first an attractive argument. We know that Rome and Constantinople were fed substantially on grain, wine and oil imported from Africa and Egypt. More than a century’s work dating and identifying the amphorae that carried the oil and wine, and the identifying of the huge harbour facilities built to receive the annual grain fleets, combine with the evidence of the Codex Theodosianus and the Codex Iustinianus to make this clear. The annona’s impact on the wider economy is at first sight obvious enough too. Throughout late antiquity, pottery types carried in the state-subsidised ships that transported these foodstuffs appear to dominate Mediterranean markets. Private shipping and alternative centres of production could not compete. Such a system fed first Rome and then Constantinople effectively, but it starved other centres of investment and diverted more natural patterns of exchange. When crisis struck the annona system in the early seventh century, there was no lively network of private and inter-regional trade to take its place.41 Twenty years ago this was a more convincing argument. Since then, however much remains obscure, it has become clear enough in the Levant and Egypt that the late antique economy was much more various and complex than once seemed the case. Whereas identifiable late antique pottery once meant in effect imported red-slip wares and amphorae, now a number of regional pottery types have come to light; and significantly, where the data has been collected, it is these regional types that dominate the ceramic assemblages. As usual, we know less about Anatolia and the Balkans, but the identification of Sagalassos ware should serve as a warning that there are likely to be many centres of regional and local late antique pottery production still to be discovered. To talk of ‘an economy dominated by annona transport’ may be overstating the case.42 The second and more recent suggestion, that we should be looking at the economic impact of state salary payments, has been made by Jairus Banaji and Peter Sarris and it fits more easily with the emerging picture of a complex late antique economy, where inter-regional and local trade was as important as what was carried on the annona trunk routes. Compared with its first- to third-century predecessor, the late antique empire was characterised by far greater numbers of official posts. The late antique aristocracy, particularly in the east, were holders of imperial office. The monetary economy of the earlier empire had been based on silver; that of the late antique empire was based on gold, specifically the solidus, struck from 309 onwards at seventytwo to the pound. It was a high-value and stable currency, and was used to pay official salaries. A middle-ranking official in the praetorian prefecture in Constantinople could earn 1,000 solidi a year and aristocratic fortunes were calculated in tens of thousands of solidi.43 To put this in context, the number of copper folleis to the solidus varied during the sixth century between 180 and 480, the former a short-term revaluation in the wake of the plague, the latter being the mode.44 Bare survival was possible on a couple of folleis a day; a workman might earn between five and twelve folleis a day. A solidus might buy a pig, three solidi a donkey, fifteen a camel, thirty a skilled slave. Forty solidi was the dowry for the daughter of an army veteran living in a Palestinian village. It might cost 400 solidi to build a village church.45 By the sixth century, the papyri evidence shows that much of the Nile valley was organised into great estates owned by this gold-rich aristocracy, who ran them as commercial enterprises, producing primarily for the market. That in turn fuelled further commercial activity and economic growth. All the indications are that late antique Egypt was highly monetarised and market-oriented.46 The papyri are effectively unique to Egypt, but the archaeological evidence cited above for late antique prosperity suggests that the Egyptian pattern held true for the sixth-century east as a whole. If this interpretation is right – and it seems to fit the evidence better than the alternatives – the state’s contribution to the success of the late antique economy was threefold. It provided a stable gold currency; it organised the fiscal system that underpinned the stability of that currency; and through official salaries it channelled wealth into the hands of the aristocracy, whose activities in turn fuelled prosperity. Such an economy could have ridden out climate change, dust-veil events and plagues – and the Egyptian evidence suggests strongly that it did – but it would have been vulnerable to political and military crisis, and that was what it faced in the seventh century.