In an empire which minted a stable silver coinage, the drahm, throughout most of its history, the continuing resort to land-grants in return for military service calls for an explanation. The drahm was the only denomination in constant circulation, raising the question whether such a simple economic system can be described as a truly advanced monetary economy.Gold dinars were issued occasionally – not, it seems, for purposes of monetary circulation, but rather in commemoration of solemn events. Bronze change seems to have been issued only very intermittently, perhaps in response to specific demands, as atMerv; the volume progressively decreased, posing problems for the mechanics of everyday economic exchanges.50 The assumption that Arsacid copper coinage was still used in many parts of the Sasanian kingdom is unconvincing,51 and the conclusion must be that much economic activity was based on barter. This situation explains a good deal about the Sasanian system of taxation before the beginning of the sixth century. It was based on crop-sharing, the exaction of agricultural produce proportionate to annual yield, as assessed by royal tax-collectors on the spot, and levied in kind. In addition, a poll tax was imposed on most subjects, which may largely have been paid in money, though part was perhaps commuted to goods. The system was inefficient and wasteful, especially with regard to the land tax; it was subject to frequent fluctuations, and allowed little scope for advance financial planning. The necessity of waiting for the tax-collector with the crops untouched in the field or on the tree meant that some might be damaged or destroyed before being enjoyed by farmers or the shah. Only lands held directly by the shah could be taxed in this manner effectively, but even on royal domains the avarice of corrupt tax-assessors will have hampered collection.52 Towards the end of the fifth century, the burden of taxation on the peasantry seems to have become increasingly oppressive: the complex relations with the Hephthalite khanate, looming in the east, resulted in heavy demands at a time when recurrent famines compelled shahs to grant occasional – and somewhat measly – tax relief. This oppression contributed significantly to the popularity of Mazdak, a heretical Zoroastrian priest, who advocated the economic equality of all human beings and regarded the higher classes of the Sasanian kingdom as the worst enemies of his doctrines. For some time he managed to enlist the support of Shah Kavad I himself: Kavad appears to have used this movement precisely in order to humble his recalcitrant nobility.53 When eventually he turned his back on the movement and allowed his son to put it down, the battered nobles needed royal support to recuperate and regain a fraction of their former grandeur. They were obviously in no position to form a viable opposition to the one serious attempt to introduce a tax reform in the Sasanian realm, begun apparently towards the end of Kavad’s reign (531) and continued by his son Khusro I.54 On the basis of a general land survey, a new system for exacting the land tax was devised. Fixed rates of tax were imposed on agricultural land according to its size and according to the kind of crops raised. The tax was calculated in drahms, although at least some was probably still levied in kind, calculated according to the current value of the produce in drahms. This new system, efficiently applied, would enable a monarch to anticipate incomes and budget expenses. It might be seen as harsh on the peasantry, primarily because the fixed drahm rates apparently disregarded fluctuations in agricultural yield caused by drought, other natural calamities or war. But this is to ignore the best testimony about the reform: if a distinction is drawn between the reform’s institution and operation in Khusro’s reign, and the way it subsequently worked, the system appears reasonably efficient and fair. It considerably augmented crown revenues, but also included a mechanism for constant revision, making tax rebates and remissions possible when and where necessary. The fiscal reform was accompanied by agricultural reform. Dispossessed farmers were restored to their lands, financial help was available to enable them to restart cultivation, and a mechanism was instituted to assist farms affected by natural disasters. The overall result should have been to maintain a system of small farms that could be taxed easily, and to prevent the growth of huge estates whose powerful owners might accumulate privileges and immunities, and obstruct effective taxation. Khusro’s reform was meant to have a lasting impact on Sasanian military organisation by providing the shah with a standing army of crack units of horsemen (asavaran), under his direct command and permanently at his disposal, who received a salary, at least while on foreign campaign. This body of palace guards was recruited from among young nobles, as well as the country gentry who wished to start a military career. On the frontiers, troops recruited from the nomadic periphery, such as Turks, as well as from semi-independent enclaves within the empire – for example, Daylam in the mountainous region of Gilan – might be employed to repel invasions or hold them up until the mobile crack units arrived. Khusro’s system appears to have enjoyed moderate success for a few decades, until the difficulties that beset the Sasanian monarchy exposed its weaknesses. In the fiscal area, its proper functioning depended on internal stability, external security and continuing financial prosperity, backed up by revenues other than the land and poll taxes – such as taxes on international trade, especially the silk trade, booty from foreign wars, tribute and diplomatic subsidies. These supplementary sources of income were necessary to ensure the smooth running of the control mechanism that was integral to Khusro’s system. However, its stability as a whole depended too much on a delicate balance which only a very powerful monarch could maintain at the best of times, and in the vast Sasanian monarchy, with its long frontiers, it was exposed to the dangers that threatened the empire itself. Growing military commitments increased the financial demands and pressure on tax-payers, thereby threatening the system; if central government lost effective control, abuse and corruption might swamp arrangements. A neglected source which appears reliable on this issue – the Sirat Anushirwan, embedded in Ibn Misqawayh’s Tajarib al-umam – indicates that towards the end of his reign, Khusro struggled to keep his system functioning. 55 The control mechanism proved to be as susceptible to corruption as the taxation machinery it was supposed to regulate. Furthermore, the strained relations between soldier and civilian, especially in the remoter zones, took their toll. In effect, the shah could restrain only those soldiers under his direct command from despoiling the rural tax-payers, as is shown by the restrictions imposed by Hormizd IV on a journey to Media. It is probable, however, that even during the last days of his father many of the cavalrymen no longer owed direct allegiance to the shah, and had reverted to being retainers of the great, virtually independent landlords. A brief glance at the aftermath of Khusro’s military reforms may help us to understand what happened. The fragility of the financial arrangements underpinning the standing army militated against enduring success for Khusro’s reforms. If, as suggested above, the Sasanian economy was never fully monetarised, the need to provide for the army’s everyday needs, at times mostly in goods, will have encouraged the reintroduction of enfeoffment as the standard military contract, even among the lower ranks. Following a short period when Khusro made serious attempts to sustain his new standing army, even in his own lifetime the asavaran increasingly reverted to an enfeoffed estate, despite such fiefs’ tendency to become hereditary and the consequent problems of alienation.56 Khusro’s reforms were, at best, of such limited duration and impact that their scope and intent might be questioned. From the royal perspective, the higher nobility posed even more serious problems than the cavalrymen. The Mazdakite revolt and its aftermath made possible a feudal system more directly dependent on the shah than ever before. The nobility restored by Khusro was firmly beholden to the shah, so there could be no doubts about the origin of its estates or the nature of the services it owed the crown. But the nobility soon returned to its former position of power. The notion that the supreme military commanders and ministers of state were now salaried civil servants is contradicted by the limited evidence available. Thus, for example, Khusro’s nominees as spahbads – the four supreme military commands he created to supersede the old office of the artestaransalar – can only have been mighty territorial lords from the start, as the very territorial nature of their command suggests. The same goes for the marzbans, the commanders of the frontier provinces. The supposition that direct dependence on Khusro as restorer and benefactor would make his nobility more tractable and obedient to the shah in the long term is not sustainable, in view of the role played by the nobility under subsequent reigns, quite apart from the revolts in Khusro’s first decade. Bahram Chobin of the noble house of Mihran, the first serious pretender outside the royal house since the establishment of the Sasanian dynasty, was supported by many disgruntled nobles. Khusro II overcame him in 591 with great difficulty, and only with the expensive support of the Byzantine emperor Maurice.57 Later, the Sasanian monarchy was rocked by other major revolts, such as those of Bistam and Bindoe – Khusro’s relatives and allies turned foes – and of his powerful general, Shahrvaraz, who was to depose his grandson Ardashir III (628–9) and claim the throne.58 By the time of the Arab conquest local rulers, especially in the east and in the Caspian provinces, had become virtually independent. The same is indicated by the confused Arabic traditions concerning Yemen after its conquest by the Persians in the last decade of Khusro I’s reign. The growing independence of the great landlords meant that sooner or later they would inevitably control not only their own retinues of fighting men but also independent taxation in their domains. Thus, for example, according to Dinawari, the future rebel Bistam, upon his nomination as governor by Khusro, instituted taxation in the territories under his rule (Khorasan Qumis,Gurgan andTabaristan) and in the process remitted half of the tax.59 Other potentates, not in direct or prospective revolt against the shah, may have acted less openly but may not have been impelled by the requirements of war propaganda to be so generous. Under Khusro II, oriental sources record impressive data about royal revenues, which might suggest that the machinery devised by Khusro I was still operating smoothly, and that Khusro II made even better use of it than his grandfather.60 But the full narrative of al-Tabari gives a different impression: the revenueswere not the product of regular taxation and should be explained in part by the influx of booty from Byzantine territories (the rich spoils of Alexandria and Jerusalem), and in part by extreme measures of extortion.61 It was primarily as an efficient operator of the taxation machinery that Khusro’s Nestorian finance minister (vastaryoshansalar), Yazdin, endeared himself to his lord; the favourable Khuzistan chronicle insists on the vast amounts of money that he sent to the treasury from the sunrise of one day to the sunrise of the next.62 Such extortions seem to have involved not only an unbearable burden on tax-payers in the royal domain but also an attempt to reintroduce direct royal taxation in the domains of grandees, who had by now come to regard this as a blatant encroachment upon their privileges: the nobles proved ultimately to be his downfall. Thus Khusro II’s riches cannot be attributed to the tax reforms of Khusro I. The last decades of the Sasanian dynasty are the story of a chain of violent upheavals, exposing all the inherent weaknesses of the huge empire. The reforms of Khusro I did constitute a serious attempt to cope with these weaknesses and to re-establish the shah’s position on a firmer basis. They failed in the long run because they strove to superimpose the framework of a fully centralised state, with a salaried civil bureaucracy and army, financed by an efficient and easily manageable taxation apparatus, on a realm which proved too weak to bear these heavy burdens. The political and military organisation of its vast territories was too flimsy, the economic infrastructure too primitive, and the social structure hidebound by traditions that could not be easily transformed. Khusro’s own conservatism was a characteristic reflection of these traditions, for it was Khusro who did much to restore the battered nobility to its traditional powers after the Mazdakite interlude. Warfare had always been the primary activity of the Sasanian state, but even by its own standards the last century of its existence witnessed a sustained intensity of campaigning that may have weakened the structures of society. After war broke out against Justin I in 527, there were only twenty-eight years of formal peace with Byzantium until the conclusive victory of Heraclius in 628 – and this is to ignore the recurrent tensions enmeshing the Arab satellites of the rival empires, Sasanian involvement in the affairs of the Arabian peninsula and the struggle to maintain control in Caucasian principalities such as Suania and Albania. We know much less about the sequence of campaigns on the north-eastern frontier, but these were probably more debilitating. Khusro’s apparent triumph over the Hephthalites in the 550s was only achieved through alliance with the rising Turkish confederation, which now replaced the Hephthalites as Persia’s neighbours and soon constituted a far more powerful threat during the 570s and 580s.63 No less than Justinian, Khusro was repeatedly involved in wars on more than one front, and the expenses of eastern campaigning probably proved much heavier than the gains from spoils, ransoms and payments stipulated in his treaties with Byzantium. The success of the state depended ultimately on the character and reputation of the shah, and there was a recurrent danger that such a personal monarchy would experience bouts of severe dynastic competition: thus, the long reigns of Shapur I and Shapur II were both followed by shorter periods of instability. This danger may have been increased in the sixth century by the withdrawal of Persian shahs from regular active participation in warfare, a move which fundamentally changed the nature of royal legitimation. Early rulers from the house of Sasan had demonstrated divine favour for their rule through personal victories, but the successors of Khusro I relied on others to win their wars.64 From the royal perspective, legitimacy ran in the family, but the nobility and armies might prefer to give their loyalty to a successful commander such as the non-Sasanian Bahram Chobin or Shahrvaraz. The existence of substantial minority religious groups, Jews as well as Christians, allowed an established ruler to secure his position by balancing their different claims against the majority Zoroastrians. But it also meant that a rival could promote himself by seeking the support of one particular group: Bahram Chobin is known for his links with the Jews. In spite of the attempted reforms of Khusro, the Sasanian state remained a fairly simple structure in which much economic and military power rested with the feudal nobility. Royal authority was bolstered by a supremacy of patronage, but this presupposed regular inflows of wealth for redistribution. Wars against the empire provided considerable short-term gains, and Byzantine peace payments under the ‘perpetual peace’ (532) and the fifty-year peace (562) were also important, but it is impossible to calculate how much of this wealth drained eastwards, almost immediately, to the Hephthalites or the Turks. The monetarised heartland of the Sasanian state (as of its Achaemenid antecedent) lay in the rich agricultural lands of Mesopotamia and lower Iraq, areas susceptible to attack from the west, and it seems to have been impossible to increase their tax revenues in the long term. It is ironic that the most successful Sasanian conqueror, Khusro II, must also bear responsibility for the monarchy’s subsequent rapid collapse. In the first decade of his reign, his status as a virtual puppet of Constantinople must have contributed to support for the long-running rebellion of Bistam in the east.65 The overthrow of his patron Maurice in 602 gave Khusro an opportunity to assert his independence, and the disorganisation of Byzantine defences, particularly during the civil war between supporters of Phocas and Heraclius in 609–11, permitted Khusro to transform a sequence of traditional lucrative frontier campaigns into a massive expansionist thrust towards the west. But whereas a war of pillage replenished royal coffers, the annexation of territories reduced the inflow of funds and meant that the newly acquired resources had to be devoted to maintaining troops in remote regions. Furthermore, Khusro’s successful armies had little direct contact with their distant monarch, being tied more closely to their victorious commanders; as a result, the soldiers of Shahrvaraz supported their general when he was threatened by the shah. In the 620s, Heraclius’ campaigns into the heart of Persia exposed the fragility of Khusro’s achievements, prompting a palace coup that introduced the most severe bout of dynastic instability the Sasanian state had ever known. The return of booty to the Byzantines together with the destruction caused by campaigns in Mesopotamia left the monarchy short of wealth and prestige at the very moment when the Arabs started to raid across the Euphrates. Yazdgard III was forced to abandon Iraq in 638–9 and thereafter lacked the resources and reputation to challenge the new Islamic superpower. The Iranian nobility abandoned the Sasanians and transferred their allegiance to the Muslim rulers, who offered stability, while the rural majority went on paying their taxes – to support a new elite.