AN UNDERDEVELOPED ECONOMY 77An argument of Hopkins^11 runs as follows: A ship of 400 tonnes can be
calculated, with the aid of comparative evidence, to have cost 250–400,000
sesterces to build and a further 185,000 sesterces to load with wheat. The
Romans are known to have built ships of 250–450 tonnes from the fi rst
century BC. Only the rich could have afforded to build and run such ships.
Those who invested in them must have included the richest men in the
empire, the Roman elite. This is the argument, in bare summary form.
Now, in our period, such large ships (most Roman ships were smaller)
were used for the transport of goods, principally wheat, to Rome, from
Egypt and north Africa. Rome was a special case. Emperors attracted
shippers into the service of the food- supply ( annona ) with tax and insurance
benefi ts, benefi ts that were available, however, only to bulk suppliers. Most
of the grain imported in this way, we may suppose, was state- owned, having
the status of tax or rent. The extent of profi t over and above that allowed
for in the contract (an unknown) would have depended on the amount of
(subsidized) trade a shipper was able to carry on alongside his function of
transporter of state goods. As for upper- class participation: senators might
have been involved, though the class of men who had the means to invest as
individuals or groups was much larger than the senatorial order. But in any
case, the involvement of rich investors, whatever their status, would tell us
more about the importance and character of ‘administered trade’ (as earlier
defi ned), than the scale of trade in the sense of commercial exchange in the
Roman world. In sum, the implications of Hopkins’ argument require
further exploration. However, unlike the conventional method of text- by-
text analysis, it is of a kind to raise the level and the tempo of the debate.
In a second, more elaborate, argument, Hopkins asserts for the period of
the Principate that the exaction of money taxes in central provinces of the
empire for expenditure on the frontiers stimulated a large volume of long-
distance trade, as taxpayers sold produce to raise cash. This, the imposition
of money taxes, is one of the mechanisms by which the monetary economy
of Roman Italy spread to other areas of the empire, in Crawford’s recent
account; Crawford adduces also the spending power of soldiers paid in coin
and the need of ambitious local aristocrats to raise cash to spend in Rome.
These hypotheses force an examination of the relative importance of money
taxes as opposed to taxes in kind, the character of the army supply system,
the nature of army pay, the political horizons of local aristocrats and their
strategies as landlords. In comparison with the Hopkins/Crawford model,
we envisage a more signifi cant role for taxes and requisitions in kind in
army supply, lower spending capacity among soldiers, who received little of
their pay as cash (see Chapter Five), lower political horizons among local
aristocrats (and therefore less adjustment of their economic behaviour) and,
in general, less penetrating monetization of local economies.^12
A third argument^13 relates to the central issue of demand. The bulk
movement of essential commodities beyond the local market, whether basic
foodstuffs such as grain, wine, olive oil and salt, or other essentials such as